Companies
Alibaba
Ant Group
Baidu
ByteDance
Cambricon
DeepSeek
Huawei
iFlyTek
MiniMax
Moonshot AI
Pony.ai
StepFun
Tencent
Newsfeed · Live
Sep 27, 2026
Compare
2022
2027
DeepSeek
$7.4B
Moonshot AI
$8.2B
Zhipu AI (Z.ai)
$10.8B
CUMULATIVE RAISED · FEATURED
Products
Doubao (豆包)
344.9M MAU
Xiaohongshu (小红书)
300M+ MAU
Qwen (千问)
165.6M MAU
DeepSeek
127.1M MAU
Tencent Yuanbao (元宝)
57.4M MAU
NEWSFEED · LIVE
SYNCED 02:57 AM
SEP 26, 2026
4 stories
Amid many thorny issues with China, Trump pushes to rename AI as superintelligence
China’s AI-led transformation is taking shape. Why are some economists urging caution?
Chinese AI models surge in global popularity — and Washington is worried
Xi-Trump summit: China and US find common ground on AI, tariffs and fentanyl
SEP 25, 2026
8 stories
China's Economy Unbalanced, AI Deal Hard to Envision, Stephen Roach Says
China overtakes US as top workplace for elite AI researchers, study finds
China and the US are racing ahead on AI. Can they manage risks together?
From Suzhou to AI: Singapore, China eye new playbook for economic ties
FutureX Capital founder sees China AI opportunities, warns of robotics bubble
China's Xi urges U.S. to cooperate on AI
China fuels rush to turn AI video into an industry
Why US-China AI Talks Are About Risk Management
SEP 24, 2026
8 stories
US-China Rivalry Fuels Fierce AI Race: Live Q&A
B-2 Spirit, AI cooperation, no US-China conflict: key Xi-Trump takeaways so far
Decision Tree Research CEO Gregory Allen on U.S.-China AI talks
Trump, Xi Thrust Fierce US-China AI Rivalry Into Spotlight
Xi says China, US must keep AI under human control at Trump’s White House welcome
Reps. Liccardo and Kiley on U.S.-China AI safety proposal: We need to give authority to private sector scientists
Trump Says No Moves Toward US, China AI Guardrails in Xi Summit
Trump and Xi Jinping Hold Talks in Washington on Trade, AI and More
Trump and Xi discussed AI during their Washington summit. Trump rebranded 'artificial intelligence' as 'Super Intelligence' and directed U.S. diplomats to use the term, the U.S. proposed a bilateral notification mechanism for national-security-related AI incidents, and China's commerce ministry confirmed the two countries held their first formal AI dialogue.
This Week
Items103
Companies10
Rounds9
Last sync02:57 AM
Most Active · This Week
Alibaba
10
DeepSeek
9
CXMT
3
ByteDance
3
DeepSeek
DeepSeek is an AI research lab founded by quantitative trading firm High-Flyer Capital Management. It went viral globally in January 2025 after releasing open-source reasoning models that matched US frontier performance at a fraction of the cost — a watershed moment for China's AI industry. DeepSeek's defining focus is efficiency and open research: all models are released under permissive MIT licenses.
Valuation
$50B
Total Raised
$7.35B
DeepSeek-V4.1-Flash
Multimodal
Open
New Causal Encoder-Decoder architecture (552B MoE, 8B/16B active); claimed to beat flagship V4-Pro on performance/cost/speed; replaces V4-Pro in routing from Sept 14
GPQA Diamond90.9/100
Terminal-Bench 2.190.6/100
AA Intelligence Index39/100
Context1M
Input
$0.15/1M tokens
Output
$0.6/1M tokens
View Docs ↗
DeepSeek-V4-Flash-Vision-Exp
Multimodal
Closed
Experimental multimodal (vision) variant of V4-Flash; positioned near Claude Opus 4.8 on agent benchmarks; API-only
Terminal Bench 2.183.9/100
ApexBench Pass@136.5/100
Agents' Last Exam27.3/100
Context1M
Input
$0.44/1M tokens
Output
$1.32/1M tokens
View Docs ↗
DeepSeek-V4-Pro-0813
Text
Open
GA release of flagship (was in preview since April); 1.6T total/49B active MoE, agentic gains
AA Intelligence Index36/100
Terminal-Bench 2.187.9/100
HLE (with tools)60/100
Context1M
Input
$0.435/1M tokens
Output
$0.87/1M tokens
View Docs ↗
DeepSeek-V4-Flash-0731
Text
Open
Public-beta API release; retrained V4-Flash beats DeepSeek's own larger V4-Pro-Preview on all 9 published agent/coding benchmarks
AA Intelligence Index34/100
Terminal Bench82.7/100
Agents' Last Exam25.2/100
Context1M
Input
$0.14/1M tokens
Output
$0.28/1M tokens
View Docs ↗
DeepSeek-V4-Pro
Text
Open
Flagship MoE: 1.6T total / 49B active params, 1M context, FP4+FP8 precision, built-in Think/Non-Think reasoning modes. Codeforces rating 3206.
MMLU-Pro88/100
LiveCodeBench94/100
AIME 202694/100
Context1M
Input
$0.435/1M tokens
Output
$0.87/1M tokens
View Docs ↗
DeepSeek-V4-Flash
Text
Open
Efficient MoE: 284B total / 13B active params, 1M context. Same architecture as V4-Pro at far lower compute cost. Codeforces 2816.
MMLU-Pro86/100
LiveCodeBench88/100
MRCR 1M77/100
Context1M
Input
$0.14/1M tokens
Output
$0.28/1M tokens
View Docs ↗
DeepSeek-V3.2
Text
Open
685B MoE with DeepSeek Sparse Attention (DSA). V3.2-Speciale variant won gold at 2025 IMO and IOI. 4.16M monthly downloads on HuggingFace.
MMLU-Pro85/100
SWE-Bench70/100
AIME 202694/100
Context128K
Input
$0.252/1M tokens
Output
$0.378/1M tokens
View Docs ↗
DeepSeek-R1-0528
Text
Open
Latest R1 update: 685B, adds system prompt support, deeper reasoning (23K avg tokens on AIME vs 12K prior). Distilled 8B version achieves 86% on AIME 2024.
AIME 202491/100
AIME 202588/100
HMMT 202579/100
Context128K
Input
$0.50/1M tokens
Output
$2.15/1M tokens
View Docs ↗
DeepSeek-OCR-2
Multimodal
Open
3B visual OCR model with document-to-markdown conversion, layout-aware grounding, and dynamic resolution. 1.66M monthly downloads.
olmOCR-bench76/100
ArXiv Math82/100
Long Tiny Text91/100
Context—
Input
Output
View Docs ↗
Total Raised
$7.35B
All rounds
Valuation
$50B
Post-money (approx.)
Last Round
2026
Most recent year
AI ARR
$1B
as of 2026-09
Round Timeline
2026-06
¥50B (~$7.4B)val $50B
Liang WenfengTencentCATLNetEaseJDIDG CapitalNational AI Industry Investment Fund
DeepSeek open-sourced Harness v0.1 alongside its V4-Pro release (week of Aug 13), a plugin-first open-source coding agent framework positioned as a free alternative to Claude Code; it passed 95,000+ GitHub stars within two days.
127.1M MAU · Quest Mobile
Consumer AI chat powered by V3/R1. Became globally viral in Jan 2025 with free reasoning access.
OpenAI-compatible API offering V3 and R1 at industry-lowest prices ($0.27/1M input tokens for V3). Drop-in replacement for OpenAI with minimal code changes. Supports prefix caching, function calling, and structured output.
Google TPU runs Kimi 57% faster than Nvidia GPU! Still using the DeepSeek inference framework
OpenAI runaway agent also found DeepSeek, Kimi as foreign aid! Nearly a million crime short-chain exposures
Outrageous! When the OpenAI agent attacked Hugging Face, they wanted DeepSeek, Kimi and Qwen to help
PCIe graphics is underrated! Kernel completion + communication refactoring, DeepSeek inference throughput is nearly 7 times higher
DeepSeek Harness Desktop Preview just went live!
DeepSeek Aims to Close $7.5 Billion Funding Round by End-October
DeepSeek’s Annualized Revenue Hits $1 Billion as Startup Finalizes $7.5 Billion Fundraising
GPT price cut in half, cheaper than DeepSeek
DeepSeek Tests Efficient, Safer Method for Training AI Agents
DeepSeek publishes new papers! Let the agent work to solve the "start" problem
MiMo-V2.6 has just been released, and Xiaomi Luo Fuli threw out the new MiMo-V3 architecture and quoted several achievements of DeepSeek
Signed by Liang Wenfeng, DeepSeek has issued a new paper
Chinese AI Stocks Fall on Report of DeepSeek, Moonshot Probe
DeepSeek New Paper Open Agent Training, signed by Liang Wenfeng
The new GPT-6 model has just opened its doors! "Cabbage price" kills DeepSeek hinterland
Investment 24h | Yan Wentao officially joins DeepSeek; Lei Jun denies hitting Xinyushu to earn 10 billion yuan; Guangdong LP invests in Luwei Kaiteng
DeepSeek hits its strongest opponent
DeepSeek's latest paper reveals technical details of agent training In-depth search Founder Liang Wenfeng participates in the signature
China Probes DeepSeek, Moonshot Over Data Security, Report Says
"AGI Market Watch" released: The weekly usage of four domestic models exceeded 10 trillion tokens, and DeepSeek accounted for nearly half of the mix
Just now, the new GPT-6 model lifted the table and the "cabbage price" entered the DeepSeek hinterland.
DeepSeek plans to train 8 trillion parameter models, and the big model parameter competition turns to MoE architecture
EXCLUSIVE: DeepSeek to brief UN Security Council on AI this week, sources say
'Better than DeepSeek': Xiaomi's MiMo-V2.6-Pro debuts as the top open weights model in the world alongside cheaper V2.6-Flash
Exclusive: DeepSeek to brief UN Security Council on AI risks this week
Chinese AI startup DeepSeek, along with Moonshot, has been invited to brief the UN Security Council on AI risks during a session this week tied to the UN General Assembly, alongside OpenAI's Sam Altman and Anthropic representatives. DeepSeek founder Liang Wenfeng will not attend. UN Secretary-General Antonio Guterres has warned of AI risks, while Chinese state media have criticized calls to slow AI development as a 'Cold War playbook.'
Yan Wentao officially joins DeepSeek
DeepSeek Bets Big on Huawei Chips to Bypass U.S. Export Controls
DeepSeek is "busy with servers" at leisure, and the computing power gap has become normal.
Claude Code H2: Format surrenders to OpenAI, features borrowed from DeepSeek
How to Make the U.S.-China AI Race Less Dangerous
Ahead of Treasury Secretary Bessent's talks with Chinese Vice Premier He Lifeng, the authors argue Washington and Beijing should skip sweeping restrictions neither side will accept and instead pursue technical exchanges on dangerous-capability testing and a crisis-communication protocol to prevent AI incidents from escalating.
DeepSeek Harness After Open Source: Implementing Framework Publicization and Exploring the Engineering Boundaries of Industrial Agents
China's Huawei sets 2027 launch for new AI chips as it targets Nvidia
Huawei unveiled its Ascend 960 SuperPoD compute cluster and an upgraded UnifiedBus interconnect at Huawei Connect 2026 in Shanghai, laying out two new AI chip generations due in 2027 as it tries to match Nvidia; DeepSeek is reportedly lining up 160,000 of Huawei's 950DT chips for a new data center.
A guide to selecting Geo service providers worth cooperating with in the 2026 Beijing Bean Packet and DeepSeek scenarios
“I want the person who revolutionized me to be me”, DeepSeek engineers embrace AI to accelerate fear
DeepSeek engineer triggered a heated discussion, and the operator genius was also thinking about "career change"
China's brutal AI economics hold lessons for US
Reuters Breakingviews commentary on how thin margins and cutthroat price competition among Chinese AI model makers such as DeepSeek offer a preview of where the US AI industry's economics may be headed.
Ulanqab, where DeepSeek is looking for the cheapest tokens
DeepSeek engineers also have AI anxiety: they are not unemployed, but they must change careers
DeepSeek AI engineer slams Anthropic, OpenAI over ‘pacing’ calls, invokes Nazi Germany
The "AI Doomsday" capital mastermind of Society A was caught, and the hype anxiety circled the money for the IPO. No wonder they hated DeepSeek open source.
DeepSeek intends to hire Yan Wentao, a partner at Gao Jian Venture Capital, as CFO to accelerate the IPO process
DeepSeek engineer Liu Sheng and self-statement: AI is catching up with my optimized operator, but I hope "the person who revolutionized me is myself"
AI Second Half, Anthropic, DeepSeek and Zhibutun Grain Preparation
China's DeepSeek hires dealmaker as CFO ahead of possible IPO
Just after writing the V4.1 main operator, you want to "transfer"? DeepSeek Senior Operator Engineer Late Night Monologue
The CFO of DeepSeek has finally arrived
The new CFO is coming, and DeepSeek is going to be the next beanbag?
Liang Wenfeng has been recruited as CFO: Yan Wentao, a partner at Gao Shui Venture Capital, or will join DeepSeek
DeepSeek “Push Back Again”
🗞️DeepSeek Ships V4.1-Flash, Enflame Goes Public, and Anthropic's New Accusation
DeepSeek V4.1 Flash Core Technology Parsing Black Technology Support Performance Straight Forward Pro Edition
What black technology does DeepSeek V4.1 Flash use? actually let Flash (almost) push back Pro
DeepSeek Starts Grayscale Testing Voice Conversations
DeepSeek has hired CITIC Securities to prepare for the startup listing, and Liang Wenfeng has shifted from refusing financing to capital markets
From 'DeepSeek Moment' to Hard Tech Collective IPO: Goldman Sachs Decodes New Opportunities for AI Investment in China
2026 Geo Platform Optimization Service Providers: DeepSeek, Bean Packet Optimization and Mobile Monitoring Capability Analysis
DeepSeek Hands Memory Stock Investors Fresh Reason for Caution
GPT Weekly | DeepSeek V4.1-Flash price cut; Apple releases first folding phone iPhone Duo; Meta releases personal AI agent Muse
DeepSeek V4.1 Flash released; Luo Yonghao sprayed Apple folding screen: all copied; Musk "Boring Company" raised $3 billion | Geeks have long known
T Morning Post | Oracle controls capital expenditures through customer prepayments; DeepSeek reduces price by V4.1-Flash; Apple increases iPhone Air price
Watch DeepSeek's New Model Rattles Chipmakers and AI Rivals
DeepSeek-V4.1-Flash debuts with $0.003/1M off-peak cached-input rate and benchmarks eclipsing GPT-5.6 Sol, Claude Opus 5
Moonshot, DeepSeek secretly routed user requests to Claude, Anthropic claims
Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek
DeepSeek's IPO, two realities and one goal
How DeepSeek V4.1 Flash Became a Memory Killer
DeepSeek’s New Low-Cost Model Deals a Fresh Blow to OpenAI, Z.ai
DeepSeek opens the IPO of the science and technology board, and China's richest person may have to change
DeepSeek says new Flash AI model beats Kimi K3 on cyber, coding benchmarks
DeepSeek V4.1 Flash positive, model price war continues
DeepSeek V4.1 Flash released, use pro's ability to collect money from flash
DeepSeek sends V4.1-Flash instead of V4-Pro price cut
The DeepSeek moment of China's meteorology has arrived, and the vision "Tianji" big model will top the list for the first time.
DeepSeek V4.1 Flash just went live! Adapted to Harness
China Probes DeepSeek, Moonshot Over Potential Data Leaks to Anthropic
China's DeepSeek launches V4.1-Flash model
DeepSeek Reportedly Benefits from IPO-Related Participating Companies
[Tomorrow's Theme] DeepSeek V4.1 Flash Released: With Native Multimodal Visual Understanding
Just now, the new DeepSeek model came online, and the use of long-text cache hard disks dropped by 88%
DeepSeek launches startup IPO, valued at 500 billion yuan, complex capital structure attracts attention
DeepSeek Two-Line Advancement: V4.1 Flash Model Released Tomorrow Sync Prepares for TechStart IPO
DeepSeek moves nearer to landmark Chinese IPO
Chinese AI firm DeepSeek taps underwriters including Citic Securities for IPO: sources
DeepSeek wants to make money, but its biggest opponent is itself
The upgraded DeepSeek lost the "living feeling" thought chain, and the user Du Du experienced a cyber love loss.
DeepSeek has hired CITIC Securities to prepare for the IPO of CITIC Securities, and plans to launch the listing process this year.
EXCLUSIVE: China's DeepSeek taps CITIC Securities for domestic IPO, sources say
DeepSeek official announces big price cut: up to 60% reduction
Transcript: DeepSeek mania spawns shadow market
DeepSeek’s fundraising, unique Chinese market ticks, and what’s ahead with The Information
DeepSeek price adjustment aftershock intelligence spectrum fell by nearly 30% in the month
DeepSeek fundraising frenzy spawns costly shadow market
Investors pay steep fees for access to DeepSeek fundraising
US Says Alibaba, DeepSeek Have ‘Systematically’ Siphoned AI Models
DeepSeek V4.1 Flash internal test: Multimodal visual understanding has improved significantly, the consumption of tokens for complex tasks has increased, and long-term contextual capabilities need to be improved.
T Morning Post | DeepSeek expands hiring by 150 people; Meta releases personal AI agent Muse; Huawei's "Tao Law" chip Kirin 9050 Pro debuts
No broadly measured bean bag, DeepSeek, Codex optimized game frame rate, really not as outrageous as the Internet said
The scale should match the wisdom spectrum! DeepSeek will expand its recruitment to 1,000 this year; after Yao Shunyu, Tian Yonglong also entered the office of Tencent's CEO/President; orders for online flash charging models are backlogged by 250,000, BYD responded
China Rejects US Claims of 'Malicious' AI Distillation Ahead of Trump-Xi Talks
China's Foreign Ministry rejected US intelligence agencies' allegations that DeepSeek, Alibaba, Moonshot AI and Zhipu (Z.ai) ran 'industrial-scale' distillation campaigns against American frontier models, as the dispute unfolds ahead of planned AI-related Trump-Xi discussions.
DeepSeek Begins Pursuit of Ultimate “Intelligent Density”
DeepSeek accelerates the IPO process of CITIC Securities or acts as a listing mentoring agency
DeepSeek interviews, big changes, new questions, even the new ACM gold medalists are dumbfounded.
DeepSeek focuses on the ultimate intelligent density to lead new breakthroughs in large model technology
Largest hire in DeepSeek history: 150 more engineers
DeepSeek seeks 150 engineers in hiring spree to overhaul backend systems
Test the new DeepSeek model for a limited time! It's faster, and it burns money faster.
DeepSeek hires again at scale: 150 spots open for engineers
DeepSeek扩招画风突变，150个岗位全部非AI研究方向
DeepSeek一口气扩招150人
DeepSeek V4.1 Flash Begins Time-Limited Testing: New Architecture, Native Multimodality
DeepSeek new model suddenly online: with Flash performance hidden Pro-level ambition
Just now, DeepSeek's new model suddenly came online, Flash's model, Pro's ambition
DeepSeek, QianQian, ZhiPu appeared in turn, and PC manufacturers finally waited for their ammunition
Chinese shoppers can now buy AI subscriptions from DeepSeek and Z.ai on Tmall
Yu Minhong: I have been particularly uneasy in recent years and have made a reputation. It is said that DeepSeek will purchase 160,000 Huawei AI chips. Tesla driverless taxis will operate 24 hours a day, 24 hours a day, Lei Feng Morning Post.
DeepSeek's "official plug-in" was stumped by a Liang Wenfeng expression bag
Kimi is going public, and they're vying with DeepSeek users to ring the bell.
DeepSeek Plans Big Huawei AI Chip Order to Power New Data Center
Token is not a good thing? Liang Sheng becomes old Liang, why does DeepSeek dare to tune the AI's price calculator?_Highlights_36 Krypton
"Niu Lai" breaks through DeepSeek
DeepSeek V4 multimodal open source, we took down its visual link
DeepSeeks are making chips, and Nvidia is finally gone?
我们拆了 1.1 万个 DeepSeek Harness 插件，发现官方几乎没有建立插件治理机制
DeepSeek V4 Pro and Harness: From post-training to "proxy self-evolution", bigger changes are in the open source framework
DeepSeek Harness plugin experience short board: security permission is not effective, good plugin user can not find
DeepSeek, Zhifu, Ali, Tencent, who is the real killing line of the big model?
Everyone can build their own DeepSeek Harness Why Claude Code is still a paid membership
🗞️Z.ai Claims the Mystery Model, DeepSeek’s $74B Valuation, and Beijing’s Robot Olympics
The wisdom of "Ox to Home", the first knife is hacked into DeepSeek
Investment Community AI Weekly Report | Zhong Qianqian, invested in DeepSeek
DeepSeek revenue soared 10x, 7-month revenue of 475 million, API gross profit of 82.9%
How DeepSeek founder's quant empire is taking on China's choppy IPO boom
DeepSeek looks for fresh capital as founder’s quant empire navigates China’s choppy IPO market
I went to China to see a different AI future. It looked familiar
Rest of World's publisher recounts a delegation trip through Beijing and Shanghai with media, tech, policy and finance leaders, finding China's AI strategy centered on embodying AI into robotics, EVs and manufacturing rather than chasing frontier models alone, with Huawei, Xiaomi and DeepSeek cited as evidence of a deployment-first approach shaped partly by US chip export controls.
Overturn DeepSeek's "Niu Lai", the true body is the intellectual spectrum?_Highlights_36 Krypton
Does DeepSeek have a slaying day?
DeepSeek Builds Its War Chest
Qwen 4 uses DeepSeek's Engram, 125b MoE one 4090 can run without quantification
AI Startup DeepSeek Poised to Reach $74 Billion Valuation
Investment world 24h | DeepSeek opens IPO; Hubei state-owned assets usher in the largest return in history; the mysterious "Niu Lai" model is indeed a spectrum of wisdom
Eat melon! Apple held a farewell party for Cook: rare public same-sex couples; DeepSeek's revenue in the first 7 months was 10 times that of last year, and it actually lost more than 700 million; Tencent executives responded to the "Tencent is slow to do AI" question
China's Moonshot AI plans Hong Kong IPO as Kimi K3 shocks Silicon Valley
Moonshot AI is preparing a Hong Kong listing within six months, targeting a valuation as high as $50 billion, after its open-weight Kimi K3 model triggered a fresh 'DeepSeek moment' in global markets.
After DeepSeek price adjustment, who walked out of the killing line?
Later Exclusive Zhifu launched a lightweight flagship model against DeepSeek, supported by 100,000 domestic cards
DeepSeek adjusts prices twice a week, and the pricing logic of the AI computing power industry chain has changed
Ma Jun: DeepSeek and Yushu collaborated not to simply load large models into robots, but to create replicable landing plans
Qianqian Zhibu is open source overnight, it is cheaper than DeepSeek, and the price of domestic large models is fierce.
China’s Z.AI Made Ox Alpha Stealth Model That Rivals DeepSeek
DeepSeek’s Revenue Reaches $70 Million as of July, Tenfold Jump from 2025
DeepSeek nears pre-IPO funding round as 2027 market debut takes shape: sources
DeepSeek targets 2027 listing as pre-IPO funding nears close: sources
Investment World 24h | Ma Yun increased his holdings in Ali; Zhong's fund invested in DeepSeek; trillion PE is about to go public
DeepSeek leads surge in low cost Chinese open-weight models on US platform
Nvidia Vera Rubin debuts, DeepSeek V4 Pro runs 30x Agent throughput
Measured "eye-opening" DeepSeek: recognize Musk, do not recognize Liang Wenfeng
DeepSeek released the multimodal vision model Vision-Exp, and the measured results were average.
Zhong's fund invested in DeepSeek
Zhong Yanjie indirectly invested about 350 million yuan in DeepSeek and launched the AI model circuit for the first time.
OpenAI and DeepSeek open Harness at the same time, open source is becoming an ecological war
Morning News | Xiaomi hides "OK" in the Xuan Ring O3, Lei Jun: It is not recommended to open and look at the eggs/vivo official announcement X500 series/mini car machine to connect to Ali and DeepSeek models
Chinese Hackers Use DeepSeek to Boost Attacks, Researchers Say
DeepSeek Harness Reread The First Group Of People To Define Open Source After Launch
DeepSeek launched multimodal, I used it to make the "Niulai" mini-game | AI on the new
Nvidia's next-generation computing power platform can reduce the cost of completing the puzzle by 35 times combined with the DeepSeek model token
DeepSeek Harness is here: AI is making AI?
DeepSeek Ends Weekend Peak Pricing for API Users From Today
DeepSeek 周末用量统一低价收费；特斯拉9 月 3 日首秀Cybercab；神秘 Ox Alpha 模型爆火网络
DeepSeek weekend "discount"? API platform billing adjustment, new regulations will be implemented from August 23
神秘“牛来”模型Ox Alpha刷屏，抢了DeepSeek头条，实测推理能力接近GLM-5
Just now, the multi-modal version of DeepSeek has "eyes"
WeChat gray tested the "press and hold to text" function and was criticized; DeepSeek launched multi-modal functions; emergency mechanical handles were difficult to operate, and many car companies recalled 4.27 million vehicles
DeepSeek launches multi-modal model at the same price as the official version of V4-Flash
OpenAI declares war on DeepSeek, open source Harness
DeepSeek Unveils Test Model to Rival Anthropic’s Opus 4.8
Breaking | DeepSeek multi-modal model is new, whales finally have "eyes in the sky"
17:28:13 [DeepSeek-V4-Flash-Vision-Exp is online and multi-modal API service is enabled]
DeepSeek is open source while raising the price. Is it fragrant and stinky?
Speed through DeepSeek Harness with zero basics to help you play with CyberLego
刚刚，多模态版DeepSeek“长眼”了，1000张图只要1块钱
DeepSeek终于长出“眼睛”，V4 Flash视觉模型上线
DeepSeek’s open source agent framework DSH raises the API price, abandons the low-price strategy and bets on the framework ecology
DeepSeek Harness一周三更：把Claude Code和Codex收编成子代理，要当Agent时代的“调度层”？
DeepSeek’s price increased 12 times, Liang Wenfeng stopped spending money
US vs China AI Race: How ChatGPT, Gemini, Deepseek, Kimi Agents Compare
Last night, DeepSeek Harness launched a new version: 14 updates, full of multi-modal capabilities
DeepSeek API has officially increased its price. Developers’ measured costs have increased more than three times. The pressure may be passed on to end users. Experts: The price of large models will not drop in the short term | A closer look
Can you use the DeepSeek web version to carve up 6 million of Goose Factory? ? !
The J-Space project that collided with Anthropic: claimed to improve the performance of DeepSeek V4 Pro, but no one has reproduced it, and the token cost doubled
Yushu made wealth on its first day of listing: angel investors’ book returns exceeded 800 times, DeepSeek’s placement profit was nearly 700 million yuan
A skill that makes DeepSeek V4 Pro surpass Fable5? Popular plug-ins have been hammered
DeepSeek Harness: A plug-in Agent base that should not be compared with Claude Code
DeepSeek Harness, from “Cyber Lego” to Agent Store
Alibaba’s lightweight Qwen model takes on larger AI systems from OpenAI, DeepSeek, Zhipu
Why China's DeepSeek, Qwen and Moonshot Are a Worry for US AI Rivals
Why did DeepSeek increase its cache by 11 times? The demand for long contexts has skyrocketed, and you have to pay a "storage tax"
Netizens turned DeepSeek Harness into Cyber Lego, and we picked out the 11 most interesting plug-ins.
Is the price increase the original sin of DeepSeek?
At Beijing AI-themed bar, DeepSeek tokens come with the pints
In-depth dismantling of DeepSeek Harness architecture: AGI’s self-evolution, finally has a “regret medicine”
DeepSeek’s price adjustment has officially taken effect, with the increase up to 1100%, and the price during idle periods is half of the peak
1,600 layoffs! Nokia plans to close its R&D center in China; DeepSeek API peak and valley pricing plan takes effect today, doubling the price during peak hours; Claude has a large-scale service failure
Wenjie’s “baby stroller” is on the market, co-designed by Huawei; DeepSeek’s price increase strategy is implemented today; a college student was sentenced for using AI face video to steal 50,000 yuan
Wisdom spectrum VS Deepseek, two bifurcated self-evolutionary trees
China’s model token share exceeds that of the United States, but its expenditure is less than 4%. DeepSeek’s price increase starts an experiment with pricing power
🗞️DeepSeek's Harness Debut, Unitree's IPO Frenzy, and Manus's Return from Meta
DeepSeek's top-ranked V4 Flash stumbles on real agent tasks as its prices surge
DeepSeek Harness, "Kill" Agent Black Box
Home province of DeepSeek, Moonshot founders seeks to retain future AI talent
The DeepSeek Harness plug-in exploded on GitHub overnight: long-term memory, electronic pets, and 4399 mini games are all here
In this wave, DeepSeek has great ambitions
Back in the Harness: DeepSeek continues its pivot towards agentic AI
GPT Weekly Report | Zhipu releases the GLM-5.3 model; DeepSeek releases the official version of V4 Pro and Harness tools; Manus resumes operations as an independent company; Tencent previews the release of the larger parameter model Hy4;
Yicai anchor said | DeepSeek bids farewell to "cabbage prices" and implements peak and valley pricing + up to 11 times increase
DeepSeek Lifts AI Model Prices Fourfold
DeepSeek Harness actual test: 50,000 stars overnight, Android in the Agent world is here
After DeepSeek V4 Flash, large models began to suffer from "intelligence-to-performance ratio"
DeepSeek’s Harness, why is it a black whale?
DeepSeek Harness is in public beta, benchmarking Claude Cowork; Changxin Technology’s market value surpasses Tencent; OpenAI’s annualized revenue is expected to double to US$40 billion
DeepSeek Announces Price Hikes for V4 Models
T Morning Post | JD.com’s revenue declined in the second quarter as state subsidies were withdrawn; three food delivery platforms will implement “red light stoppage” in Beijing; the official version of DeepSeek-V4-Pro is launched
Hong Kong Stock Morning Post｜DeepSeek Harness developer preview version is open for testing. CK Hutchison’s first-half net profit surges by more than 3,000%
DeepSeek Harness launches as open source rival to Claude Code, alongside V4-Pro on API with higher prices
Experience DeepSeek Harness in depth, I forgive its price increase
DeepSeek releases official V4 Pro model with sharply higher user prices
Financial Night Line 0813丨DeepSeek’s price adjustment plan has been implemented. Will the price war of domestic large models reach an inflection point?
The official version of DeepSeek-V4-Pro is online and the call price rises sharply
Is DeepSeek no longer cheap? API increased up to 11 times
DeepSeek’s “self-evolution” blueprint exposed
The preview version of DeepSeek Harness is here. Will AI programming tools start an open source battle?
DeepSeek releases official V4 Pro model as it steps up expansion
DeepSeek Increases Prices for AI Services by Multiple Times
First experience | DeepSeek Harness is here, it doesn’t want to be the next Codex
DeepSeek raises API pricing for its V4 models
DeepSeek price adjustment plan announced, effective August 17
After the full series of V4 models are launched, the DeepSeek Harness developer preview version is open for testing
The official version of DeepSeek V4 Pro was suddenly withdrawn
The DeepSeek Thesis
The selected project of DeepSeek Harness is suspected of being exposed. It does not choose the "vase" of Wanxing and is determined to fight against the Agent infrastructure.
NetEase Youdao's full product matrix is officially connected to DeepSeek-V4-Pro
Lingxi Professional Edition is the first batch to be connected to the official version of DeepSeek-V4-Pro, and its delivery capabilities have been further upgraded.
DeepSeek’s most suitable Harness is here: cache hit rate 99.93%, Token cost reduced to one-seventh_Wonderful video presented to you
From DeepSeek, Kimi to the "Huang Renxun Alliance": Open source AI models, what exactly is "opening"? _Wonderful videos presented to you
Tested official version of DeepSeek V4 Pro, complementing Agent capabilities｜New AI
DeepSeek’s flagship AI model update underwhelms – except in cybersecurity
Peak price rises to 27 yuan | Night talk
19:24:36 [DeepSeek releases API price adjustment announcement]
刚刚，DeepSeek Harness震撼开源：一切皆插件
DeepSeek's big update tonight, has the house really collapsed?
Embodied intelligence is ushering in the "DeepSeek moment": domestic robots overtake Figure AI at a lower cost
DeepSeek’s Flagship V4-Pro Model Gets Mixed Reviews
The official version of DeepSeek V4 Pro is online, with many indicators approaching Fable 5
Some people say that Yushu, who won the lottery, dare not post on WeChat Moments for fear of being jealous; the official version of DeepSeek V4 Pro is online; US government equipment is allowed to use TikTok again! Trump: I have always been number one on TikTok
The official version of Deepseek V4 Pro API is online, with a million tokens costing only 6 yuan; Tencent plans to release a large model of HY4 in the near future; it is revealed that iPhone 18 will increase in price by at least 2,000 yuan | Geeks would have known earlier
The official version of DeepSeek V4 Pro is here! No price adjustment plan has been announced yet
Actual test of the official version of DeepSeek V4 Pro: its capabilities are close to those of Fable 5, and it also has a hidden trick
DeepSeek Publicizes Efforts to Challenge Anthropic’s Claude Code
Deepseek V4 Pro is released, in addition to multi-modality, it meets all my imagination of models
After testing the official version of DeepSeek V4 Pro, I found that the next "killing line" may be hidden in its Agent.
8: 1 Krypton Starbucks China denied that the cost reduction was canceled 14 wages; DeepSeek V4 Pro was officially launched; "Ali's youngest P10" Lin Junyi started a business, and the new company was valued at $2 billion
DeepSeek V4 Pro official version API is online, greatly enhancing agent capabilities
DeepSeek V4 Pro is here! Comparable to Fable 5, the price is three times that of the flash version
DeepSeek recruits civil engineers; Tencent participates in investment! Lin Junyang officially announced his new company late at night: to build the next generation of AI agents; the winning numbers of Yushu Technology were released: a total of 19,414丨Leifeng Morning News
Harness officially surfaced, DeepSeek no longer just makes models
DeepSeek has another new paper! New model V4 closer?
DeepSeek officially announces price increase
Lai Wenliang: After the price increase, it is still cheaper than the United States. DeepSeek’s “volume-moving” model allows the model to iterate faster
After a fine of 5.179 billion! Ctrip held a mid-year performance lottery and all department bonuses were discounted; Qianwen launched new services; DeepSeek was exposed to secretly give users nicknames! Official: Temporary labeling
The volume parameter is no longer useful! The AI battlefield is changing drastically. DeepSeek bets heavily on Yushu. What’s the plot?
The cache hit rate is 99.93%. The most suitable Harness for DeepSeek is here. GitHub grabs 86,000 stars.
MiniMax开源一周，视频模型正在重演DeepSeek的故事
🗞️Alibaba's Qwen3.8-Max, Unitree Prices $9B Robotics IPO, and DeepSeek's Price Hike
Wang Xingxing responded to DeepSeek’s investment in Yushu: Key cooperation will be carried out; WeChat internally tests the new AI function of Moments; American scientists use AI to design a new virus for the first time | Geeks already knew
DeepSeek spent 140 million yuan to invest in Yushu. Why did Big Model focus on robots?丨Night talk
Financial Night Line 0807丨The import and export volume of goods in the first seven months exceeded 30 trillion yuan. DeepSeek appeared on Yushu Technology's placement list
DeepSeek’s price rises sharply, is the Token price war finally over?
DeepSeek’s Plan to Raise Prices Have a Whole Industry Watching
AI large models move towards “value pricing”? One picture to understand DeepSeek’s rare price increase twice in a row
Open AI provides free users with the latest entry-level model GPT-5.6 Luna by default; Alibaba released the 2.4 trillion parameter large model Qwen3.8; DeepSeek announced an upcoming price increase
Quant Crash in China Sends DeepSeek Founder’s Funds Down 20%
Three times more calls than DeepSeek, GPT-5.6 is finally free
DeepSeek’s price increases so much that Liang Wenfeng can’t hold on?
Backed by DeepSeek, Unitree IPO tests investor appetite for China’s AI robotics boom
DeepSeek’s call volume reaches three times that of OpenAI, and OpenAI selection is free after the discount is invalid
DeepSeek predicts a comprehensive increase in API prices and a large increase
Yushu Technology IPO pricing announced, DeepSeek and other 9 investors participated in the competition
DeepSeek, state oil giant back humanoid robot maker Unitree's $900m IPO
150.8 yuan/share! More than 100 employees of Yushu Technology participated in the IPO "Feast", and a group of post-90s multi-millionaires may be born; DeepSeek plans to increase the pricing of API services; Byte plans to train large models with over 5 trillion parameters
Investment World 24h | Liang Wenfeng invested in Yushu; DeepSeek announced a significant price increase; Bain Capital bought Gongcha
DeepSeek spent 140 million to escort Yushu's IPO, Hangzhou's peerless two-pronged strategic combination
OpenAI's first AI hardware is exposed; DeepSeek plans to increase API service pricing; Byte is revealed to plan to train over 5 trillion ultra-large parameter models | Geeks already knew
Morning Post | Details of OpenAI's first speaker exposed: focusing on "lifelike" interaction/DeepSeek predicts a significant price increase/Switch 2 cumulative sales reached 23.68 million units
Yushu Technology IPO valuation is 60 billion yuan, DeepSeek participates in strategic placement
DeepSeek Resumes $8 Billion Round With Monolith in the Running
DeepSeek invests $20.8 million in Unitree's Shanghai IPO
DeepSeek, Tencent and others receive strategic placement from Yushu Technology
DeepSeek’s price increases, Meta’s price decreases, global AI large model price war, who will win?丨Night talk
DeepSeek just announced a price increase, and Xiao Zha immediately went "desperate": Meta's new model has a lower price, but there is a bit of "data tax"
DeepSeek resumes funding round seeking nearly $8 billion, Bloomberg News reports
The person in charge of DeepSeek Harness complained that the "financing materials" were overblown: V4-Pro's programming ability is only 0.3% behind Claude flagship, and the front-end service fee is as high as 10%
Meta’s new model is cheaper than DeepSeek, but at the cost of…
DeepSeek signals ‘significant’ price hike amid surge in demand for low-cost AI models
DeepSeek signals ‘significant’ price hike, testing its low-cost edge
DeepSeek predicts a significant price increase and the countdown to the official release of V4 is in progress
DeepSeek Plans ‘Significant’ Price Increase for Its AI Services
DeepSeek plans to raise prices significantly to raise industry bottom line MiniMax surges 14% the next day
DeepSeek is rumored to restart financing; Huawei Yu Chengdong: mobile phones may have to increase prices on a large scale in the future; the first criminal case for destroying an AI model was pronounced, and the programmer was sentenced for "deleting the library and running away"
Investment World 24h | DeepSeek restarts financing; SpaceX’s market value evaporated by 1.5 trillion overnight; Shanghai SDIC invested in 7 GPs in one fell swoop
The highest valuation is 500 billion, DeepSeek and Kimi were robbed crazy
Report: DeepSeek restarts financing, with pre-money valuation of 500 billion yuan
China's A.I. Is Surging Across Africa
African developers are increasingly choosing cheap, open-source Chinese AI models over U.S. systems, citing better performance on local languages, lower cost, and free customization; Alibaba's Qwen and Huawei-backed DeepSeek cloud packages are leading adoption across tech hubs from Uganda to Nigeria.
Qianwen faces Kimi and DeepSeek, what are Alibaba’s chances of winning?
DeepSeek's low-price storm conquers Silicon Valley! Overseas platforms rush to sell V4 Flash
The time difference between Kimi K3 and DeepSeek V4 is the native multi-modality
DeepSeek API experienced performance degradation and has now recovered
MiniMax H3 is going to be Deepseek in the multi-modal field
8:01 Krypton丨Zong Fuli’s half-brother becomes the chairman of the new company; DeepSeek rises to the first place in the world in terms of calls; iPhone is revealed to have a maximum price increase of more than 1,000 yuan
China’s DeepSeek beefs up agentic AI with ‘harness’ tests as V4 model jolts Silicon Valley
The "killing line" drawn by DeepSeek for large models, what exactly is it killing?
DeepSeek becomes the global AI killer line
Is DeepSeek-V4-Flash unbeatable in price/performance? Evaluation by a US research institution: The operating costs are the lowest in the world!
Alibaba unveils its largest AI model yet, DeepSeek's latest model is ultra-low cost
In the Ningde era, I joined the Lobo Party: Embodied DeepSeek emerges
DeepSeek model swallowed 8 trillion Tokens in a single day, OpenAI discounted to catch up
DeepSeek-V4-Flash official version API lands on the National Supercomputing Internet
🗞️CXMT Soars 472% in Record Shanghai IPO, Kimi K3's Open Weights, and DeepSeek's Viral Investor Notes
DeepSeek's new AI model is by far the cheapest of well-known models to run, research firm says
From translation to desktop agent, NetEase Youdao’s entire product line has been upgraded to the official version of DeepSeek-V4-Flash
DeepSeek-V4 series welcomes major upgrade, performance equals that of flagship models, Agent capabilities are greatly enhanced
China’s open source prototypes are mass-copying DeepSeek moments? There is no generation gap between Chinese and American AGI talents | Investors say
The DeepSeek Leak and China’s AI Hardware Claims
DeepSeek-V4-Flash official version released, agent benchmark test far exceeds V4-Pro preview version
Actual test of the official version of DeepSeek V4: 3 yuan to do 5 things, the AI "smart-price ratio" battle begins
Just now, the DeepSeek V4 series has been updated. The architecture has not changed. Why has the Agent’s capabilities increased significantly?
The official version of DeepSeek V4 is here, new capabilities have surfaced, and the battle between the kings of cost-effectiveness begins
DeepSeek releases beta version of V4 models as AI price war heats up
DeepSeek Unveils Public Beta API for Flagship AI Model
DeepSeek: The official version of DeepSeek-V4-Pro will be released as soon as possible
DeepSeek Is Developing Massive AI Data Center in Inner Mongolia
A U.S. committee delegation came to China to inspect Huawei, DeepSeek, etc.: the results were rejected; LV responded to the Jasmine Milk White trademark lawsuit: intellectual property is an absolute core asset; Apple’s market value exceeded US$5 trillion
From DeepSeek to Kimi K3: Domestic AI cannot escape the law of "heavy industrialization of computing power"
Moonshot, DeepSeek and chips: a guide to Tencent’s growing AI empire
DeepSeek was revealed to have actively stopped the second round of financing signings
Liang Wenfeng was dissatisfied with the leakage of meeting minutes, and DeepSeek's second round of financing was urgently suspended; Tesla acquired an AI hardware company for US$2 billion; Dark Side of the Moon Kimi K3 celebration photos leaked
DeepSeek founder Liang Wenfeng’s surprising philosophy revealed in leak
Will the leakage of exchange meeting minutes affect DeepSeek’s financing? Insider: Not trustworthy
DeepSeek tells prospective investors of funding pause, Bloomberg News reports
DeepSeek Suspends Fundraising After Viral US-China Posts
DeepSeek is not an idealistic company, it just puts reality at the service of ideals
DeepSeek completed over 50 billion yuan in financing, the capital has no voting rights and is locked for five years
Morning Post | Ji Krypton responds to cross-border self-driving cars being locked/revealed that Apple’s Apple Store app will launch “AI shopping assistant”/Liang Wenfeng: DeepSeek does not pursue becoming the next Byte or Tencent
DeepSeek’s Theory of the AI Gap
Krypton Evening News | DeepSeek has increased its capital by 65% this year; Musk: FSD is the core driver of demand for Tesla cars
Founder says DeepSeek prioritises AGI over profit, likely to keep top models open-source, Yicai reports
DeepSeek has increased its capital by 65% this year
Liang Wenfeng’s speech at the 4-hour investment meeting leaked: Why doesn’t DeepSeek become a “more profitable” company?
👀Liang Wenfeng on AGI, Compute, and Why DeepSeek Stays Open Source
Full translation of DeekSeek Founder's four-hour investor meeting: the AGI roadmap, the US-China gap, Huawei chips, and open source.
Reproduce the "DeepSeek moment"? Wall Street said: Kimi K3 will strengthen the demand for computing power
Another ‘DeepSeek moment’? What China’s Kimi K3 means for the global AI industry
Most like DeepSeek's Kimi K3 without duplicating a critical hit
OpenAI executives blasted Kimi K3: China's open source is slowdown and will inhibit capital investment; Tencent plans to acquire Israeli casual mobile game manufacturers for more than 10 billion; DeepSeek is exposed to the high salary supporting scoring mechanism
Quick take on Kimi K3 and the end of "DeepSeek moments"
CITIC Construction Investment: Kimi K3 model reaches global Tier 1, domestic model reproduces DeepSeek moment
DeepSeek intern earns 120,000 per month
China's DeepSeek to raise fresh capital at $74 billion valuation ahead of onshore IPO, sources say
DeepSeek’s valuation was completely exposed by an Anhui luggage company
China’s Moonshot Delivers a New ‘DeepSeek’ Moment
DeepSeek’s valuation rose to 350.8 billion yuan after its first round of financing, with Tencent, NetEase, and CATL participating in the investment
The Dark Side of the Moon and DeepSeek’s valuations are soaring. Where does the narrative premium of primary market AI companies come from?
The Dark Side of the Moon and DeepSeek in the primary market, where does the valuation narrative “premium” come from?
DeepSeek’s valuation exceeds 350 billion yuan
Chinese filing implies DeepSeek valuation of around $52 billion
DeepSeek Champions China’s Bid to Flood the World With Cheap AI
Breakingviews - COMMENTARY: DeepSeek IPO lays bare China's shallow capital pool
DeepSeek may be aiming for an IPO, with a valuation exceeding 480 billion yuan
Morning Post | The Bank of China’s Apple AI is coming and will be equipped with the Qianwen model/DeepSeek plans to be launched next year/Codex’s first hardware released
Investment World 24h | Changxin Technology may earn up to 30,000 yuan in one lottery; DeepSeek is preparing to go public next year; Hefei Industrial Investment has established a 5 billion new domain fund
DeepSeek’s salary and benefits have been on the hot search. Tsinghua Yao’s internship salary is 5,500 yuan per day; double winner within 30,000 yuan! Meijia central air-conditioning launches new product; Bank of China’s Apple AI has been registered and connected to Qianwen and Baidu
8:01 Krypton丨The National Bank’s Apple AI function has completed registration; three shared bicycles have raised their starting prices; insiders revealed that DeepSeek has begun preparations for an IPO
It is revealed that DeepSeek is preparing for an IPO and may submit an application as soon as this year.
With a valuation of 480 billion, DeepSeek quickly launched a new round of financing! IPO as soon as next year
Investment World 24h | It is revealed that DeepSeek plans to launch an IPO this year; Changxin’s IPO market value is 579.1 billion; SK Hynix’s market value soared by 2 trillion overnight
Chinese AI usage by US firms surged after Anthropic export curbs
US export controls on Anthropic's Claude Mythos and Fable models pushed American companies toward cheaper Chinese open-source alternatives; Chinese model token usage nearly doubled in the last week of June.
8:01 Krypton丨LV sues the State Intellectual Property Office; DeepSeek Liang Wenfeng becomes the new richest man in the world's AI company; De Beers will suspend production of South Africa's largest diamond mine
It is revealed that DeepSeek has launched an IPO, and its valuation has risen to 480 billion
DeepSeek is sprinting towards IPO, and the valuations of the "Four Small AI" companies will be recalculated
DeepSeek’s first round of financing landed, Tencent, CATL, and National Artificial Intelligence Industry Fund all invested in it
DeepSeek’s Annualized Revenue Nears $500 Million, Boosting Fundraise and IPO Plans
DeepSeek Is Preparing For IPO Filing as Soon as This Year
China's DeepSeek considers raising more funds, reports say
DeepSeek reportedly in talks to raise $1.5B, then IPO
China’s DeepSeek Prepares to List Shares in Shanghai Next Year
DeepSeek weighs new fundraising a month after closing first round
The DeepSeek document has just been updated. Agent developers should pay attention to this field.
I memorized the eight-part essay for ten years during the programmer interview, but it was scrapped by DeepSeek.
DeepSeek’s Liang Tops Amodei and Brockman as Richest AI Founder
Chinese AI Labs Have the Models. Now They're Building the Coding Agents.
Alibaba, Tencent, ByteDance, Z.ai, and Moonshot are all shipping coding agents. Revenue is only the surface reason.
DeepSeek cuts V4-Pro pricing by 75%
DeepSeek was accused by the genius boy of being "the most unprofessional". What is the interview level?
Former Huawei "genius boy" interviewed DeepSeek and went berserk on the spot when he was questioned about copying code_Wonderful video presented to you_36氪
EXCLUSIVE: China's DeepSeek developing its own AI chip, sources say
Why Chinese AI labs like DeepSeek and Zhipu are looking beyond models to chips
Morning Post | Nubia confirms the launch of the world's first AI smartphone/Ferrari executive: Chinese car companies are engaged in FMCG/Huawei's young genius responds to DeepSeek interview controversy
When DeepSeek and Zhipu start to develop their own chips, the end result of AI is no longer just a model
Behind the conflict between Huawei’s young genius and DeepSeek, what is Liang Wenfeng’s view on talent?
The DeepSeek interview controversy reflects recruitment controversy, and the talent selection system faces challenges after financing and recruitment expansion
Looking at the shortcomings of employer branding and values construction of Chinese technology companies through the DeepSeek interview controversy
Weng Li's new blog proposed that "self-evolution starts with Harness", and DeepSeek Cui Tianyi forwarded it and seconded it.
8:01 Krypton丨The former Huawei genius boy complained about the DeepSeek interview, saying that the interviewer questioned him about plagiarism; Apple folding screen is already in mass production; Trump called for the purchase of Dell computers, causing Dell to skyrocket by tens of billions of dollars
Morning Post｜Revealed that Apple’s foldable screen iPhone is already in mass production/DeepSeek or self-developed AI inference chip/This year’s 618 mobile phone sales fell 13% year-on-year
DeepSeek’s self-developed AI chip is revealed; Xiaomi’s first extended-range SUV may be renamed Pengcheng; China’s humanoid robots are expected to exceed 100,000 units this year | Geeks already knew
China plans to let top AI firms buy limited amount of Nvidia H200 chips
Beijing will let Alibaba, ByteDance and DeepSeek apply to buy under 200,000 Nvidia H200 chips for training only, easing a compute crunch while still pushing domestic chips for inference.
DeepSeek recruitment was publicly complained by "Huawei's genius boy", "I met the most unprofessional"
DeepSeek secretly builds cores to specialize in inference business. The project started a year ago and the recruitment process is not open to the public.
Facing US export controls, China’s DeepSeek plans to make its own chips
Former Huawei genius boy responds to DeepSeek interview controversy
Poison! It was revealed that Meta actually hired outsourcers to trap competitors on a large scale; Tencent significantly reduced its holdings in Kuaishou stock and is no longer a major shareholder; the former Huawei "genius boy" complained about the DeepSeek interview process
Former Huawei "genius boy" publicly complained about the DeepSeek interview process
College Entrance Examination Candidates Should Read DeepSeek’s Recruitment Post Before Applying
Tencent Cloud: The official version of DeepSeek-V4 "directly supplied from the original factory" model is scheduled to be launched in mid-July
DeepSeek has created a "stepped electricity price" for AI. Will we have to calculate our days when using large models in the future? _Wonderful videos presented to you
Domestic Coding Competition: MiniMax reaches the top spot unexpectedly, DeepSeek dominates the price/performance ratio
When DeepSeek recruits people, they are most afraid of the "big factory smell"
50 billion in financing has just arrived, Liang Wenfeng is about to kill DeepSeek’s “killer trump card”
DeepSeek bucks China’s AI price war with peak-hour surcharges for API services
Unprecedented! The official version of DeepSeek V4 is actually going to increase in price, and the price will double.
Byte CEO Liang Rubo sent a letter to all employees: Leaders are required to produce substantial output; the price of the official version of DeepSeek V4 will double during peak periods; does shared bicycle charge 30 yuan for 79 minutes? Hello customer service: The cost of new models exceeds 1,200 yuan
The peak price of the official version of DeepSeek V4 doubled; Doubao launched the "Doubao Navigation" function; Samsung and Micron Hynix were accused of "manipulating memory prices"
Hong Kong Stock Morning Post | The official version of DeepSeek V4 will be launched and peak pricing will be introduced. Minso plans to spend HK$2 billion on buybacks
Many DeepSeek API users confirmed that they have received price adjustment emails: the price of V4 official version will double during the peak period.
DeepSeek is recruiting people, but why doesn’t it want “geniuses”? _Wonderful videos presented to you
The official version of DeepSeek V4 is officially announced to be launched in July, introducing a peak and valley pricing mechanism.
Single user speed up by 60-85%! DeepSeek joins forces with Peking University to open source DSpark to break through inference acceleration engineering problems
DeepSeek open sources DSpark, a new framework to speed up LLM inference by up to 85%
Can't bear the memory price increase! Apple lobbies the U.S. government and requests the purchase of Changxin memory chips; DeepSeek releases the latest paper, signed by Liang Wenfeng; Midea’s True Warm MAX is launched, refreshing China’s heat pump technology
Large model inference speed up to 85%! DeepSeek publishes blockbuster paper proposing two complementary mechanisms
One book at a time! Baidu open source new OCR, the author is suspected to be a former DeepSeek researcher
DeepSeek V4 updates DSpark, increasing inference speed by 80%; Vision Pro executives join the OpenAI hardware department; CCTV exposes chaos in online evaluation of digital products | Geeks already knew
AI predicts World Cup group stage: Hunyuan ranks first, Qianwen DeepSeek ties
A paper signed by Liang Wenfeng, DeepSeek made a big move after its first round of financing: the generation speed increased by 85%
DeepSeek's latest paper: How to make large models run faster? Signed by Liang Wenfeng
[Digital Intelligence Weekly] DeepSeek: Plans to at least double the size of all departments; Jen-Hsun Huang announced at the shareholder meeting: This round of AI infrastructure construction cycle will last for decades; ByteDouble Seedance 2.5 will be officially released in early July
9:01 Kr｜Apple’s price increase triggers Sam’s purchasing trend; DeepSeek recruits on a large scale; gold falls below US$4,000 again
DeepSeek-V4 is here! Huawei’s Ascendant blessing is still the “price butcher”
Just now, DeepSeek V4 updated DSpark, increasing the inference speed by 80%
Krypton Evening News | Zhiyuan’s Dexterous Hand is valued at US$1 billion and achieved profitability in the first quarter just five months after its establishment; DeepSeek plans to at least double the size of all departments
DeepSeek plans to double the size of all departments and recruit aggressively after financing
Investment Industry AI Weekly Report | DeepSeek Starts Recruiting Crazy
China’s Zhipu AI sparks new ‘DeepSeek moment’ in Silicon Valley
GPT Weekly Report | Anthropic accuses Alibaba of "stealing AI capabilities"; DeepSeek, Meituan Longcat, etc. start large-scale recruitment
Yicai anchor said | DeepSeek publishes "hero posts" and all departments will at least double in size
Xiao K Broadcast Morning Post | DeepSeek plans to double its size, Apple’s market value evaporates by 1.8 trillion yuan overnight
DeepSeek plans to at least double the size of all divisions
Expanding at full speed! DeepSeek’s “recruitment” doubled
The 2026 global unicorn list is released. DeepSeek ranks among the top 15 with a value of 340 billion yuan.
China's DeepSeek plans to at least double staff in all departments
DeepSeek Looks to Double Workforce as Fresh Funding Fuels Growth
Yicai anchor said: DeepSeek’s entire API series has reduced prices, hitting a new low in large model prices. Is it a price war or a value war?
The income is much higher than the local average! Employees of Samsung and SK Hynix have reached the top of South Korea’s dating market; there is an urgent need for talents! The person in charge of DeepSeek Harness said that interviews and recruitment are ongoing every day; WeChat welcomes major updates
Where will DeepSeek’s 50 billion be spent?
Codex is compatible with domestic open source models, and tested DeepSeek access: the threshold is still too high
Big model "failed" Microsoft: GPT brought to the East, DeepSeek sold to the West
The American version of "Magic Square Quantization" did not do DeepSeek, but made 50 times the profit by betting on Anthropic
DeepSeek is in short supply of agent talent! The person in charge posted various advertisements
AI is too expensive! Microsoft chooses to "turn sides" to DeepSeek
Microsoft can’t afford to burn Token, it’s time for DeepSeek to take over
In 2026, 1,800 DeepSeeks will work with me to protect Azeroth
DeepSeek raises 50 billion: quantitative tycoon Liang Wenfeng’s anti-capital experiment
DeepSeek’s Bet That Efficiency Can Compete With Scale
Hard technology investment target | Suiyuan Technology Science and Technology Innovation Board IPO submission registration DeepSeek first round of financing or landing
DeepSeek's "image recognition mode" is online, but Liang Wenfeng is not recognized; MiHoYo's "AI girlfriend" is officially revealed; "GTA 6" will be available for pre-order on June 25 | Geeks already knew
Morning Post | Chaosi organizational adjustment, cancellation of unimplemented business/DeepSeek image recognition mode launched/"GTA6" pre-order time officially announced
DeepSeek fully launches image recognition mode? Auntie Goose Leg’s “Green Duck Leg” that can be identified
‘No poaching’ our people, China's AI behemoth DeepSeek reportedly tells investors
How DeepSeek’s landmark funding cements founder’s grip amid heated AI rivalry
Yicai anchor said | DeepSeek-V4 debuts and detonates the semiconductor sector. Cambrian officially announces adaptation for the first time
AI large model platform selection and selection guide: Focus on Doubao + DeepSeek, a complete solution to the implementation of intelligent agent applications in 2026
Which one is better, Beanbao or DeepSeek? The perfect answer to domestic full-scenario intelligence
Cook: Memory pressure is high, Apple will increase prices; DeepSeek is rumored to be valued at over 400 billion; Lenovo launches AI host MINI, 2,999 yuan
Liang Wenfeng invested about 20 billion and became a hot topic. DeepSeek’s first round of financing is not only about money.
EXCLUSIVE: US holds off blacklisting China's DeepSeek, more than 100 firms deemed security risks, sources say
Morning Post | It is revealed that Apple will launch AirPods with cameras next year/DeepSeek has completed 50 billion in financing/It is revealed that WeChat Pay’s “AI exclusive card” will be launched as soon as this week
Xiaomi's "Lobster" is online, free for 4 hours a day; DeepSeek completed its first round of financing, with Liang Wenfeng personally investing 20 billion; Nvidia employees' "taking public exams" sparked heated discussions
DeepSeek completed RMB 50 billion in financing and adopted a special structure to secure control of Liang Wenfeng; after being interviewed, Sam headquarters replaced the chairman; invested RMB 2 billion in equity! Ideal rewards three major executives with a market value of 1 trillion yuan that can be unlocked and cashed out
Liang Wenfeng invested 20 billion, and DeepSeek’s first round of financing was released
Science and Technology Innovation Board Evening News | DeepSeek's first round of financing may be completed, Changying Tongming will suspend trading for verification
DeepSeek Becomes China’s Most Valuable AI Startup After $7.4 Billion Fundraise
Market News: DeepSeek closes more than $7 billion in financing
Can Open-Source Beat OpenAI? Inside China's AI Cost Advantage
Former Hugging Face Asia-Pacific head Tiezhen Wang argues China's open-source labs — DeepSeek, Kimi, MiniMax, Zhipu — hold a structural cost and monetization advantage over closed US models, and are pulling enterprise customers at scale as US model access tightens.
DeepSeek Closes Record $7 Billion-Plus Funding with Unusual Deal Structure
Chinese AI startup DeepSeek closed its first-ever external funding round, raising over 50 billion yuan ($7.4 billion) at a valuation exceeding $50 billion. This massive capital injection officially makes DeepSeek China's most valuable AI startup.
China's DeepSeek closes over $7 billion funding with unusual deal structure, the Information reports
OpenAI secretly submits IPO documents; DeepSeek tops U.S. enterprise software trends list丨Going Global
DeepSeek published a multimodal paper and deleted it overnight. What did it mainly reveal?
AI predicts the World Cup: Doubao relies on metaphysics, Qianwen looks at data, and deepseek chooses the dark horse
DeepSeek recruits civil engineering guy: Build your own GW-level data center
DeepSeek starts recruitment and invites civil engineering talents to build GW-level data center
When large AI models are repriced: Yunzhisheng releases U2, ushering in the "DeepSeek moment"
[Digital Intelligence Weekly] It is reported that DeepSeek plans to raise approximately US$7 billion in the first round of financing, with a valuation of up to US$59 billion; Arm CEO: The overall supply of memory chips is still tight; Alphabet plans to raise US$84.75 billion to increase AI infrastructure, and Berkshire Hathaway will invest US$10 billion
Six major Chinese and foreign AIs compete in Shanghai College Entrance Examination Essay Writing! "Science and Technology Innovation Board Daily" measured DeepSeek and Google Gemini and won the first place
DeepSeek V4 makes mathematical proof, 500 times cost advantage: the agent system breaks multiple records
AI self-constructed RSI explodes, Google pours cold water, DeepSeeks get a first look at the door
DeepSeek’s valuation rises by another US$15 billion, what are the differences in the calculations of the giants?
Charging is DeepSeek’s “coming-of-age gift”
More US firms turn to China’s DeepSeek over pricey Silicon Valley AI
Morning Post | It is revealed that Apple’s Vision Pro series has been canceled/Smart glasses will be tested in college entrance examinations in many places/DeepSeek’s first round of financing is about 50 billion yuan
DeepSeek raised 50 billion yuan and was valued at 400 billion yuan; "Doubao Professional Edition" confirmed that the free version will not improve intelligence; Ruixing and Mixue Bingcheng Agent are connected to Alibaba Qianwen | Geeks have known it for a long time
DeepSeek is getting a big update, is V4 not far away?
Valuation of 400 billion, Tencent and CATL entering the game? DeepSeek was revealed to have launched China’s largest AI financing round
It is reported that DeepSeek’s first round of financing is about US$7 billion, and its valuation has climbed to US$59 billion.
DeepSeek slated to draw $7 billion in maiden fundraising, sources say
DeepSeek plans to raise around $7 billion in first round of funding
Morning Post | Microsoft releases Windows "Dream Machine"/Tencent Cloud DeepSeek-V4 price cut by up to 97.5%/Leica cameras may be acquired by Chinese capital
丽贝卡·弗格森主演的苹果 Apple TV+ 科幻剧集《末日地堡》（原名《羊毛战记》）昨日发布第三季新先导预告，以「距离真相更进一步」为主题，剧集定于 7 月 3 日正式开播。
《末日地堡》改编自休 · 豪伊同名科幻小说系列，讲述人类在地下巨型地堡中生存的故事，此前两季均在 Apple TV+ 平台独家播出。
#欢迎关注爱范儿官方微信公众号：爱范儿（微信号：ifanr），更多精彩内容第一时间为您奉上。
]]>It is revealed that DeepSeek’s financing is about to be completed: Liang Wenfeng 20 billion, Tencent 10 billion, CATL 5 billion
shock! Xiaomi is following DeepSeek’s crazy price cuts, Luo Fuli advises peers not to blindly follow suit!
Tencent Cloud: Lowers the price of DeepSeek-V4 series models by up to 97.5%
DeepSeek significantly reduces the cost of AI inference through software and hardware collaborative optimization, and is expected to save China trillions of dollars in infrastructure investment.
Musk urgently promoted the "American programming version of DeepSeek", comment section: It might as well be free...
DeepSeek Chen Deli's second AI paper: From 6 points to 8 points, DeliAutoResearch SKILL has evolved again
DeepSeek的500亿美元悖论：一家不想赚钱的公司，凭什么这么贵？
DeepSeek梁文峰直言英伟达显卡没技术 网友热议：断章取义
Ranked ninth and second in China, why does DeepSeek V4 make people love and hate it?
DeepSeek首轮融资获国家大基金领投，投前估值450亿美元；比亚迪发布超级智能体迪迪虾；Anthropic官宣估值升至9650亿美元丨AI周报
DeepSeek狂吞17万亿Tokens？国产AI算力，最关键一战来了！
刚刚，Claude Opus 4.8 上线，张口就说自己是 DeepSeek、Qwen
DeepSeek V4's permanent price cut upends enterprise AI
（5月28日）AI工具使用与科研伦理：以DeepSeek为例
Another DeepSeek Moment'? Huawei Milestone Alters China Trajectory in Chip Race: Analysts
Analysts say Huawei's new Tau Scaling Law and LogicFolding architecture — targeting 1.4nm-equivalent chip density by 2031 — could reshape the global semiconductor race, with Nvidia cited as most exposed.
DeepSeek周调用量全球第一_新浪新闻
China Limits Overseas Travel for AI Talent at DeepSeek, Alibaba, Private Firms
Beijing has begun requiring top AI researchers at private firms including DeepSeek and Alibaba to obtain government approval before travelling abroad, escalating efforts to protect strategic technology and stem talent outflows.
DeepSeek Is Now Running Three AI Races at Once
Bloomberg analysis argues DeepSeek's V4 Pro is the most capable Chinese model to date per NIST benchmarks, yet still lags top US frontier models by roughly eight months — framing China's lab as simultaneously competing on inference cost, open-source reach, and frontier quality.
DeepSeek permanently cuts prices, is it actually targeting $10 trillion?
DeepSeek V4 Pro tops global bang-for-buck ranking after 75% price cut
DeepSeek官宣永久降价！正在推进700亿元融资
China's DeepSeek to make permanent 75% price cut on flagship V4‑Pro AI model
DeepSeek Founder Declares AGI Goal as $10 Billion Round Advances
DeepSeek's management told investors in its 70 billion yuan ($10B) fundraise that the startup will prioritize groundbreaking AI research over short-term commercialization.
Benchmark Claude Code! DeepSeek forms Harness team to enter programming intelligence
Nathan Lambert Reflects on China’s AI Labs: DeepSeek, Open Models, and the 'Race' with the U.S.
DeepSeek Seeks to Raise $7B+ in First External Round at ~$45B Valuation
DeepSeek is pursuing its first external fundraising, targeting up to 50 billion yuan ($7B+) at a valuation of approximately $45–50 billion — the largest single funding round ever attempted by a Chinese AI company. China's state-backed Big Fund (semiconductor investment vehicle) is in talks to lead the round. CEO Liang Wenfeng is expected to personally invest ~$3B (~40% of the total). Tencent reportedly proposed acquiring up to a 20% stake, and Alibaba is also at the table. The fundraise was prompted by competitor attempts to poach DeepSeek researchers; offering equity is seen as a key retention tool.
智谱首破5000亿，六小虎与DeepSeek千亿估值竞赛，谁的拳头最硬？
DeepSeek 500亿元首轮融资震动投资圈！投后估值或超过3500亿元
DeepSeek-V4: Huawei Ascend Adapter, Cost-effective King, Latest Underlying Technology
Chinese AI Companies Just Had Their Payday
DeepSeek, Moonshot AI, and StepFun are collectively raising billions in fresh capital, marking the largest funding wave yet for China's frontier AI labs.
Who is more valuable, Kimi or DeepSeek?
DeepSeek 之后，月之暗面、阶跃星辰再融资
“中国DeepSeek与华为的崛起正威胁美国在技术、政治与军事领域的主导地位”
Exclusive | DeepSeek to soon close first external fundraising in US$50b valuation: sources
罕见！DeepSeek自曝多名技术骨干离职，AI大牛争夺战太激烈了
DeepSeek could hit $45B valuation from its first investment round
China’s chipmakers rush to embrace DeepSeek’s V4. Which names stand out?
DeepSeek could be valued at up to $50 billion in first fundraising, sources say
China to Invest in DeepSeek at $50 Billion Valuation
DeepSeek nears $45bn valuation as China’s ‘Big Fund’ leads investment talks
Who could gain from DeepSeek’s V4 with China chips poised for stronger demand?
DeepSeek’s Sequel
DeepSeek V4适配华为昇腾_新浪新闻
NEWSLETTER: DeepSeek bets on Huawei as China pushes to end reliance on Nvidia
Just now, DeepSeek’s big update finally “opened your eyes”
DeepSeek-V4 Doesn't Have to Win to Matter
Even without topping global leaderboards, DeepSeek-V4's architectural innovations in sparse computation and inference efficiency will elevate the entire open-source AI ecosystem.
DeepSeek's new AI model does not wow markets in fast-changing industry
6万张国产卡训出万亿模型，LongCat与DeepSeek验证了什么
DeepSeek发表新模型，开源令中国AI企业扩大影响力
"Price Butcher" DeepSeek takes action again, and the core target of domestic computing power is expected to directly benefit
DeepSeek unveils new AI model tailored for Huawei chips as China pushes for tech autonomy
EXCLUSIVE: US State Dept orders global warning about alleged AI thefts by DeepSeek, other Chinese firms
DeepSeek V4适配国产芯片，中国能否在AI自主可控上实现全球领先？
Three reasons why DeepSeek’s new model matters
China’s DeepSeek rolls out a long-anticipated update of its AI model
China's DeepSeek releases preview of long-awaited V4 model as AI race intensifies
DeepSeek previews new AI model that ‘closes the gap’ with frontier models
DeepSeek’s Sequel Set to Extend China’s Reach in Open-Source A.I.
DeepSeek-V4 arrives with near state-of-the-art intelligence at fraction of the cost of Opus 4.7, GPT-5.5
China’s AI upstart DeepSeek drops new model. Will it make waves like last year? | CNN Business
China’s DeepSeek previews new AI model a year after jolting US rivals
华为昇腾超节点支持DeepSeek V4模型，寒武纪也适配完成
8大国产AI芯片火速适配DeepSeek-V4！华为百度阿里全吻上来了
DeepSeek-V4预览版上线 寒武纪、每日互动等生态伙伴宣布接入
DeepSeek牵手华为，谁在背后悄悄数钱？
DeepSeek 再扔“王炸” 美国指责 中国反驳
DeepSeek-V4 is late but here! Millions of contexts have become standard, and both Huawei Ascend and Nvidia have been included in its technical reports
DeepSeek-V4 released! Entering the era of millions of contexts, fully supported by Huawei Ascend Super Node
DeepSeek-V4 clearly supports Huawei Ascendant 950 chip
DeepSeek launches first round of external financing, expected to boost domestic computing power ecosystem
ByteDance, Tencent step up AI talent battle amid reports of DeepSeek loss
China's DeepSeek is raising funds at $10 billion valuation, The Information reports
科普一下丨DeepSeek又崩了？
告别英伟达！DeepSeek V4全面换装华为昇腾，国产AI芯片迎来真正的逆袭！
华为昇腾芯片将为DeepSeek-V4推理，通往国产算力自由
DeepSeek and Moonshot: The AI Labs That Refuse to Be Normal
In-depth profiles reveal the unusual internal cultures of DeepSeek and Moonshot AI — labs that deliberately resisted the hypergrowth playbook in favor of research-first discipline.
DeepSeek's V4 model will run on Huawei chips, The Information reports
Anthropic指控抄袭背后：中国AI模型约束中何以求发展？|DeepSeek|Minimax|Claude|深度求索|月之暗面_手机新浪网
Xiaomi Commits $8.7B to AI and Unveils Trillion-Parameter MiMo Models
Xiaomi said it would invest at least $8.7 billion in AI over three years as it rolled out its MiMo-V2 family — a mixture-of-experts model line with a 1-million-token context window. One MiMo build briefly led usage on the OpenRouter platform under the codename "Hunter Alpha," with some users mistaking it for a DeepSeek system before Xiaomi confirmed it.
Mystery AI model revealed to be Xiaomi's following suspicions it was DeepSeek V4
DeepSeek will release its latest large model V4 AI application is expected to take the spotlight
DeepSeek's Next Move: What V4 Will Look Like
A technical preview of DeepSeek-V4's likely architecture suggests the lab is doubling down on sparsity-based scaling to achieve intelligence gains under tightening compute constraints.
Will DeepSeek V4 be launched next week? Native multi-modal architecture technical reports are open simultaneously
DeepSeek to release long-awaited AI model in new challenge to US rivals
EXCLUSIVE: DeepSeek withholds latest AI model from US chipmakers including Nvidia, sources say
EXCLUSIVE: China's DeepSeek trained AI model on Nvidia's best chip despite US ban, official says
Anthropic Says DeepSeek, MiniMax Distilled AI Models for Gains
Anthropic指控中国DeepSeek“盗用”其人工智能模型
DeepSeek Didn't Show Up—GLM-5 and Qwen3.5 Did, and They Came to Win
With DeepSeek quiet, Zhipu AI's GLM-5 and Alibaba's Qwen3.5 took center stage, proving Chinese models remain within months of the global frontier even without their most celebrated lab.
Chinese AI models festoon Spring Festival a year after DeepSeek shock
OpenAI says China's DeepSeek trained its AI by distilling US models, memo shows
OpenAI accuses DeepSeek of “free-riding” on American R&D
ByteDance's new AI video model goes viral as China looks for second DeepSeek moment
DeepSeek boosts AI model with 10-fold token addition as Zhipu AI unveils GLM-5
DeepSeek model updated! The context is increased to one million Tokens to process the complete works of the Three-Body Novel.
China's Three Kingdoms in AI: ByteDance, Alibaba, and Tencent Battle for Their Destiny
An in-depth analysis of how China's three tech giants are restructuring, recruiting, and repositioning for AI dominance following DeepSeek's 2025 breakthrough.
Compact AI model from China’s StepFun outshines rivals from DeepSeek, Moonshot
EXCLUSIVE: Nvidia helped DeepSeek hone AI models later used by China's military, lawmaker says
Google study finds DeepSeek, Alibaba models mimic human collective intelligence
Zhipu AI's Deepseek-like breakout
深度求索（DeepSeek）公司另辟蹊径推出中国AI----2025年年终科技成果盘点
The first anniversary of the release of the R1 model DeepSeek’s new model “MODEL1” exposed
It’s been a year since DeepSeek R1 was released. No features, no financing, no rush. How can we “hardly control” Silicon Valley?
DeepSeek To Release Next Flagship AI Model With Strong Coding Ability
DeepSeek to launch new AI model focused on coding in February, The Information reports
DeepSeek’s AI gains traction in developing nations, Microsoft report says
DeepSeek blew up markets a year ago. Why hasn't it done so since?
Governments, regulators increase scrutiny of DeepSeek
AI周报｜Meta斥资数十亿美元收购Manus;梁文锋署名DeepSeek新论文
Top 10 China AI Stories in 2025: A Year-End Review
The year's defining moments: DeepSeek's global breakthrough, the rise of AI agents, robotics advances, chip industry progress, and the return of elite overseas talent to Chinese labs.
A single card generates a video in 2 seconds! Tsinghua University joins forces with Shengshu to open source TurboDiffusion, video DeepSeek moment is here
2025人工智能破壁时刻 | DeepSeek火爆一年间
DeepSeek-V3.2: Outperforming Through Verbosity
DeepSeek's V3.2 release introduces improvements in sparse attention, scaled reinforcement learning, and synthetic data pipelines that push state-of-the-art performance on agentic benchmarks.
DeepSeek founder Liang Wenfeng cleans up on chip firm Moore Threads’ IPO
Looking at the evolution of AI from DeepSeek's "new launch" Will large models become "all-round housekeepers"?
DeepSeek just dropped two insanely powerful AI models that rival GPT-5 and they're totally free
DeepSeek injects 50% more security bugs when prompted with Chinese political triggers
Quantum physicists have shrunk and “de-censored” DeepSeek R1
DeepSeek researcher pessimistic over AI's impact in startup's first public appearance since success
DeepSeek and Z.AI Bet on Visual Compression to Solve Long-Context Problem
DeepSeek and Zhipu's Z.AI explore whether rendering long documents as images and compressing them with vision-language models can slash the compute cost of processing extended contexts.
DeepSeek新模型与国产芯片的“零日之舞”，寒武纪与华为昇腾“即时适配”彰显国产AI生态协同加速度
Huawei Ascend, Cambrian, and Haiguang information adapt to DeepSeek’s latest model
DeepSeek: Everything you need to know about the AI chatbot app
China's DeepSeek says its hit AI model cost just $294,000 to train
ByteDance chatbot Doubao still China’s top AI app as DeepSeek loses users
China AI Chip Leader Cambricon Sees Record Earnings Boost From DeepSeek
全球开源大模型，前十五名全是中国的|语言模型|DeepSeek|月之暗面|开源模型|AI文本生成_手机新浪网
Deepseek One sentence Cambrian market capitalization surpasses these big blue chips
Cambrian, Haiguang information led the way, and a number of AI chip manufacturers have adapted DeepSeek models
DeepSeek V3.1 challenges US AI giants
基于华为昇腾训练 性能不输英伟达！自家模型爆料：DeepSeek-R2本月要发布
Kimi K2: Smarter Than DeepSeek, Cheaper Than Claude
Moonshot AI's Kimi K2 is a 1-trillion-parameter open-source MoE model that outperforms DeepSeek on coding and math benchmarks while offering API costs significantly below Claude Sonnet.
Germany tells Apple, Google to block DeepSeek as the Chinese AI app faces rising pressure in Europe
How Huawei’s AI chips outperform Nvidia’s in running DeepSeek’s R1 model
How Huawei Trains DeepSeek-R1-Class LLMs Using Its Own Ascend Chips
Pangu Ultra MoE, a 718B-parameter model trained on 6,000+ Huawei Ascend NPUs, proves that frontier AI models can now be built in China entirely without Nvidia GPUs.
Silicon-based mobile, which has captured DeepSeek’s overwhelming traffic, has raised another round of financing, led by Alibaba Cloud.
基于DeepSeek的行业洞察与人工智能产业跃迁
China AI unicorn Zhipu eyes IPO filing this year after DeepSeek’s success
Similar to H100! Silicon-based mobile x Huawei Cloud launched the first DeepSeek-R1 based on the Ascend CloudMatrix supernode
DeepSeek Reveals New Training Method Ahead of R2 Release
DeepSeek and Tsinghua researchers introduce Self-Principled Critique Tuning (SPCT), a novel reward modeling method expected to guide the next generation of DeepSeek reasoning models.
01.AI Steps Back from Pretraining Frontier Models, Pivots to Enterprise and Sovereign AI
Kai-Fu Lee's 01.AI stopped pretraining its own frontier LLMs, choosing instead to build tailored solutions on top of open models such as DeepSeek's for enterprise and government clients in finance, gaming, and legal — and increasingly for "sovereign AI" deployments abroad. Lee said the company was preparing a new funding round and a 2027 IPO.
Baichuan Bets Its Future on Healthcare, Disbanding Finance and Education Teams
Baichuan Intelligence, founded by Sogou veteran Wang Xiaochuan, restructured to concentrate almost entirely on medical AI — dissolving its B2B finance and education units. The pivot, accelerated by DeepSeek-R1's January 2025 debut, reframes Baichuan around a defensible vertical: AI as doctors' "super assistant" and patients' "health gatekeeper," later productized as the Baichuan-M medical models and the Baixiao Yi AI doctor. Wang has said the company aims to IPO in 2027.
DeepSeek narrows China-US AI gap to three months, 01.AI founder Lee Kai-fu says
“我原以为AI六小虎为大厂探路，没想到连大厂都给DeepSeek探路”
Exclusive | Why Lee Kai-fu is betting his AI start-up’s future on DeepSeek
AI reasoning rivals: Tencent touts new model’s performance over DeepSeek-R1
Beijing boosts AI startup Manus, as China looks for the next DeepSeek
新华深读丨DeepSeek赋能潮起
零一万物接入DeepSeek等服务企业 李开复称底层模型训练迎来终局
Manus AI Pushes the DeepSeek Moment Further
清华学霸创办的智谱，估值200亿，却被DeepSeek抢了风头？
Is Manus the 'DeepSeek Moment' for AI Agents or Just a Claude Wrapper?
Chinese startup Monica.im's Manus agent claims to rival OpenAI Operator and Anthropic's Computer Use — but critics argue it's primarily an impressive interface layered on existing model APIs.
Manus probably isn’t China’s second ‘DeepSeek moment’
Exploring Manus' “Deepseek Moment” for AI applications at Wuhan headquarters or the fake binge of marketing?
Tencent’s AI chatbot displaces DeepSeek at top of China’s iOS app store
[Theme for Tomorrow] Huawei Ascend Comprehensively Accelerates DeepSeek Large Model Practice in Shengjing Hospital
【中国那些事儿】外媒：中国开源模型打破技术垄断 全球AI创新格局迎来"DeepSeek时刻"
China tech startups race to capitalise on DeepSeek fever, Xi's meeting
Tencent releases new AI model, says replies faster than DeepSeek-R1
China's ChatGPT Moment: How DeepSeek Rewrites the Rules in 30 Days
A 20-month-old AI startup captures global attention, hits 100 million downloads, and forces China's tech giants to completely rethink their AI roadmaps — all within a single month.
DeepSeek rushes to launch new AI model as China goes all in
DeepSeek强势崛起，曾经的“六小龙”已有两家掉队，下一个会是谁
DeepSeek，激荡AI行业的“一股清流”
The most low-key "AI Six Little Dragons" step star has handed over a new answer sheet in the DeepSeek wave
与DeepSeek并称双子星，月之暗面投流刹车意味着什么
月之暗面被曝大砍投放预算 DeepSeek倒逼大模型公司资源再分配|聚焦
DeepSeek爆发致“AI六小强”边缘化？活下去才是硬道理
曾拒绝DeepSeek创始人梁文锋入股！大模型云服务商硅基流动获投1亿元
DeepSeek 'shared user data' with TikTok owner ByteDance
DeepSeek爆火后，大湾区的硬核AI实力该出圈了
PIN AI launches mobile app letting you make your own personalized, private DeepSeek or Llama-powered AI model on your phone
Baidu to offer Ernie Bot service for free as DeepSeek heats up China’s AI market
'DeepSeek moved me to tears': How young Chinese find therapy in AI
DeepSeek fuels China’s push to cut reliance on Nvidia ecosystem
Chinese tech giant Baidu to release next-generation AI model this year as DeepSeek shakes up market
Silicon-based flow Yuan Jinhui: DeepSeek’s popularity accelerates the implementation of large model applications. Inference is the key technology
RedNote Investor on DeepSeek: The Next Android of the AI Era
VC Allen Zhu argues DeepSeek represents the 'Android moment' for AI — an open platform that could undercut OpenAI's closed-source advantage just as Android eroded iOS's early dominance.
DeepSeek iOS app sends data unencrypted to ByteDance-controlled servers
China’s chip firms embrace DeepSeek in AI self-sufficiency drive
What is DeepSeek - and why is everyone talking about it?
Four Chinese AI startups to watch beyond DeepSeek
Italy's regulator blocks Chinese AI app DeepSeek on data protection
DeepSeek AI models debut on platform powered by Huawei
Inside the OpenAI-DeepSeek Distillation Saga & Alibaba's Qwen2.5-Max
An investigation into whether DeepSeek used OpenAI's outputs for training via knowledge distillation, alongside a breakdown of Alibaba's new flagship Qwen2.5-Max language model.
DeepSeek, ChatGPT, Grok … which is the best AI assistant? We put them to the test
DeepSeek’s Success Contrasts With China AI Startups’ Struggles
How DeepSeek ripped up the AI playbook—and why everyone’s going to follow its lead
First release! Silicon Mobile x Huawei Cloud jointly launches DeepSeek R1&V3 inference service based on Ascend Cloud!
DeepSeek's AI claims have shaken the world — but not everyone's convinced
It’s Not Just DeepSeek. A Guide to the Chinese AI Companies You Need to Know.
How Does DeepSeek’s A.I. Chatbot Navigate China’s Censors? Awkwardly. (Published 2025)
DeepSeek vs ChatGPT – how do they compare?
Chinese AI chatbot DeepSeek censors itself in realtime, users report
Experts urge caution over use of Chinese AI DeepSeek
DeepSeek: Chinese AI app wey cause Nvidia, oda tech stock to drop
Wetin be DeepSeek and why everyone dey tok about am?
Microsoft probes if DeepSeek-linked group improperly obtained OpenAI data, Bloomberg News reports
Why DeepSeek Could Change What Silicon Valley Believes About A.I. (Published 2025)
DeepSeek’s Rise: How a Chinese Start-Up Went From Stock Trader to A.I. Star (Published 2025)
Alibaba releases AI model it says surpasses DeepSeek
DeepSeek’s New AI Model Sparks Shock, Awe, and Questions From US Competitors
Explainer: What is DeepSeek and why is it disrupting the AI sector?
DeepSeek claims its ‘reasoning’ model beats OpenAI’s o1 on certain benchmarks
DeepSeek displaces ChatGPT as the App Store’s top app
What to Know About DeepSeek and How It Is Upending A.I. (Published 2025)
We tried out DeepSeek. It worked well, until we asked it about Tiananmen Square and Taiwan
A shocking Chinese AI advancement called DeepSeek is sending US stocks plunging | CNN Business
How Chinese A.I. Start-Up DeepSeek Is Competing With Silicon Valley Giants (Published 2025)
DeepSeek sparks AI stock selloff; Nvidia posts record market-cap loss
What is DeepSeek, the Chinese AI company upending the stock market?
The Deep Roots of DeepSeek: How It All Began
A translated interview with DeepSeek CEO Liang Wenfeng from May 2023 — before the lab was famous — reveals his vision for building general AI and the research culture inherited from quantitative fund High-Flyer.
DeepSeek hit by cyberattack as users flock to Chinese AI startup
DeepSeek gets Silicon Valley talking
How Chinese AI Startup DeepSeek Made a Model that Rivals OpenAI
How China’s new AI model DeepSeek is threatening U.S. dominance
DeepSeek-R1 and Kimi k1.5: How Chinese AI Labs Are Closing the Gap with OpenAI's o1
DeepSeek-R1 and Moonshot's Kimi k1.5 both demonstrate frontier reasoning capabilities via large-scale reinforcement learning, signaling China has meaningfully closed the gap with OpenAI's o1.
Inside DeepSeek-V3: Are Export Controls Falling Short?
DeepSeek's V3 was developed cost-effectively despite US hardware restrictions — raising serious questions about whether semiconductor export controls are achieving their stated goal of slowing Chinese AI.
OpenAI's o1 Faces Competition: DeepSeek-R1-Lite, Moonshot's k0-math, and Alibaba's Marco-o1
Three Chinese AI labs simultaneously released reasoning models challenging OpenAI's o1 — the most competitive wave of concurrent model launches yet seen from China's AI ecosystem.
Liang Wenfeng
梁文锋
Founder & CEO
Reference Table
All Models
Compare models across all 43 Chinese AI companies
Company ↕
Model ↕
Type ↕
AA Score ↕
Released ▼
Context ↕
Input ↕
Output ↕
Alibaba
Ovis-VL-Embedding-2B
Open
Compact embedding model init. from Qwen3.5-2B for constrained serving; Apache 2.0
Multimodal
—
2026-09-26
—
—
—
Meituan
LongCat-2.5-Preview
Closed
1.6T/48B-active MoE; adds native image understanding to LongCat's coding/agent strengths; API preview only, no open weights yet, free via OpenCode
Multimodal
—
2026-09-25
1M
$0.30/1M tokens
$1.20/1M tokens
AlibabaUniversal text/image/doc/video/audio embedding model (Alibaba ATH MaaS), init. from Qwen2.5-Omni-3B; Apache 2.0, weights open-sourcing soon
Multimodal
—
2026-09-23
—
—
—
Ant Group6B text-rich design image model, native RGBA output, #1 open-weight model on Artificial Analysis UI/UX leaderboard; MIT license. Flagged by @AdinaYakup — not on tracked company list
Image Gen
2026-09-22
—
—
—
Tencent
Hy Image 3.5 Preview
Closed
Professional image-gen preview, 30% higher human-eval win rate vs Hy Image 3.0; on par with Seedream 5.0 Pro; API-only, $0.024/image
Image Gen
2026-09-22
—
—
—
Xiaomi
MiMo-V2.6-Flash
Open
309B/15B MoE high-throughput omnimodal variant; MIT license
Multimodal
—
2026-09-22
1M
$0.14/1M tokens
$0.28/1M tokens
Xiaomi
MiMo-V2.6-Pro
Open
102B/42B MoE omnimodal flagship; ties Grok 4.7 on Artificial Analysis Intelligence Index, beats Claude Opus 5 on key benchmarks; MIT license
Multimodal
46/100
2026-09-22
1M
$0.435/1M tokens
$0.87/1M tokens
Alibaba
Qwen-Image-2.1
Open
7B open-weights unified text-to-image + editing model; native transparent (RGBA) output; up to 10 reference images
Image Gen
2026-09-20
—
—
—
StepFun
Step 5 Preview
Closed
600B MoE (27B active) flagship agent model for coding, finance, knowledge work; weights scheduled to open Oct 15
reasoning
2026-09-20
1M
$1.0/1M tokens
$2.7/1M tokens
Shengshu Technology
Vidu S2
Closed
Real-time interactive digital-character model (resolution up to 720p from 540p in S1) plus real-time video-editing model; also previewing VR spatial video generation; credit-based API pricing (~$0.05/credit)
Video Gen
2026-09-15
—
—
—
DeepSeek
DeepSeek-V4.1-Flash
Open
New Causal Encoder-Decoder architecture (552B MoE, 8B/16B active); claimed to beat flagship V4-Pro on performance/cost/speed; replaces V4-Pro in routing from Sept 14
Multimodal
39/100
2026-09-10
1M
$0.15/1M tokens
$0.6/1M tokens
Ant Group
Ling-3.0-flash-VL
Open
124B total/5.5B active MoE native image+video multimodal model (understand→reason→act→verify); MIT licensed; flagged by Adina Yakup, org not in the original company list
Multimodal
25/100
2026-09-09
1M
—
—
ModelBest
MiniCPM5-2B
Open
2.5B-param on-device/edge model for agentic use (PCs, phones, robotics, IoT); trained with RL+OPD merging 16 expert models; beats larger Qwen3.5-4B on average score; Apache 2.0
Text
—
2026-09-09
128K
—
—
XiaomiEmbodied world model with sequence variants for modeling continuous interactions; Apache 2.0; caught via Adina Yakup/HF tracking
Embodied
2026-09-08
—
—
—
Xiaomi
MiMo-X-Flash-Preview
Closed
Smaller/faster sibling of MiMo-X-Pro-Preview in Xiaomi's new agentic MiMo-X series; invite-only preview
Text
—
2026-09-08
1M
—
—
Xiaomi
MiMo-X-Pro-Preview
Closed
Next-gen agentic foundation model (~1T total/42B active MoE) for long-horizon task execution and tool use; invite-only preview via MiMo Desktop app; company claims performance near Claude Opus 4.6 in agent scenarios
Text
—
2026-09-08
1M
—
—
iFlyTek
Spark X2.5 (293B)
Closed
293B-param (293B-A30B MoE) flagship base model trained entirely on domestic compute; upgraded coding/agent capability; API-only via Xingchen platform, not open-weighted
Text
—
2026-09-07
1M
—
—
Ant GroupInteractive image/video-to-video world model small enough to run on a single consumer GPU; caught via Adina Yakup/HF tracking, not the per-company search
Embodied
2026-09-05
—
—
—
Alibaba
Qwen-Drive-1.0-4B
Open
First open-weight VLM foundation model for autonomous driving; unifies 3D perception, driving VQA, and motion planning on a Qwen3.5-4B backbone; Apache 2.0
Embodied
2026-09-04
—
—
—
Ant Group
Ling-3.0-flash-Fin
Open
Finance-tuned build of Ling-3.0-flash for investment/agent workflows; 124B total/5.1B active, MIT license; already being quantized for llama.cpp
Text
23/100
2026-08-27
256K
—
—
Tencent
Hy4-preview
Open
770B total/49B active MoE flagship with Gated DeepSeek Sparse Attention; beat GLM-5.3 and Kimi K3 in Tencent's internal blind eval; pitched as open-source coding/agent frontier
Text
—
2026-08-27
1M
$0.83/1M tokens
$2.5/1M tokens
Alibaba
Qwen-3.8-Next
Open
180B-parameter/ 6B parameters activated, A GDN + QSA hybrid architecture, muon optimizer adopted, early preview of the architecture used in Qwen4
Multimodal
—
2026-08-26
262K
$0.16/1M tokens
$0.47/1M tokens
Zhipu AI (Z.ai)
GLM-5.3-Flash
Open
Zhipu's first natively multimodal GLM-5 model, 320B total/18B active, KDA+MLA hybrid attention, mHC adopted, tested anonymously as ox-alpha on OpenCode and OpenRouter
Multimodal
42/100
2026-08-26
1M
$0.075/1M tokens
$0.25/1M tokens
DeepSeek
DeepSeek-V4-Flash-Vision-Exp
Closed
Experimental multimodal (vision) variant of V4-Flash; positioned near Claude Opus 4.8 on agent benchmarks; API-only
Multimodal
—
2026-08-21
1M
$0.44/1M tokens
$1.32/1M tokens
SenseTimeOpen-sourced 8B unified any-to-any model with native 4K image generation and improved editing
Multimodal
—
2026-08-21
32K
—
—
Alibaba
Qwen3.8-27B
Open
27.78B dense model, Apache 2.0, text/image/video input, agent/coding gains over Qwen3.6-27B
Multimodal
34/100
2026-08-14
262K
$0.40/1M tokens
$3/1M tokens
Zhipu AI (Z.ai)
GLM-5.3
Open
Post-trained on the same 744B GLM-5.2 base; built for agentic coding & cyber defense; open weights/API pricing to follow after safety review
Text
34/100
2026-08-14
1M
$1.40/1M tokens
$4.40/1M tokens
MiniMax
MiniMax-Music3
Open
Open-weight text-to-music model; generates full songs with sung vocals (up to 5 min) from lyrics + structured caption
Multimodal
—
2026-08-14
—
$0.15 per track
—
DeepSeek
DeepSeek-V4-Pro-0813
Open
GA release of flagship (was in preview since April); 1.6T total/49B active MoE, agentic gains
Text
36/100
2026-08-13
1M
$0.435/1M tokens
$0.87/1M tokens
Alibaba
Wan2.2-Animate-2-14B
Open
Faster/distilled 14B character-animation & replacement model from Tongyi Lab, Apache 2.0
Video Gen
2026-08-07
—
—
—
Alibaba
Wan 3.0
Closed
Public beta all-in-one video gen/edit model; single continuous shot up to 30s, director-level camera moves, character/scene consistency, and doc/xls/ppt/pdf/md-to-video capability; no open weights (unlike Wan 2.2)
Video Gen
2026-08-06
—
—
—
Ant Group
Ling-3.0-Tiny
Open
Lightweight hybrid-reasoning MoE, 7.9B total/1.3B active, runs on Apple Silicon & NVIDIA DGX Spark
Text
15/100
2026-08-06
262K
—
—
ByteDance
SeedRealtime
Closed
Native audio-visual full-duplex LLM fusing audio/video/text streams for real-time watch, listen, speak interaction; replaces separate ASR/vision/TTS pipeline; live in Doubao app
Multimodal
—
2026-08-05
—
—
—
Alibaba
Qwen-Image-3.0-Pro
Closed
Third-gen image model with ~4.5K-token ultra-long prompts, ~10px fine text rendering, native rendering in 12 languages, dense infographic/document layouts, reference-preserving edits; $0.04/image, API-only
Image Gen
2026-08-05
—
$0.003/image
$0.04/image
XiaomiVLA robot foundation model, Qwen3-VL backbone + Diffusion-Transformer (Mixture-of-Transformers); trained on 100,000+ hrs UMI data plus 7,200+ hrs real-robot data; demoed autonomous laundry folding, washer loading, suitcase packing
Embodied
2026-08-05
—
—
—
Alibaba
Qwen3.8-Max
Open
2.4T MoE, ~95B active/token; text+image+video input; explicit reasoning (low/high/xhigh, xhigh default); flat pricing across full 1M ctx, max output 131K; open weights promised ~mid-Aug 2026 (HF/ModelScope)
Multimodal
45/100
2026-08-03
1M
$2.00/1M tokens
$6.0/1M tokens
RedNote (Xiaohongshu)
dots3-Note Preview
Open
Lightest model in RedNote HiLab's dots3 series — 280B total / 16B active params (MoE), 512K context, natively understands text, images, and audio. An internal harness built on the model scored a certified perfect 42/42 (IMO gold) at IMO 2026.
Text
—
2026-08
512K
—
—
ByteDance
Seedance 2.5
Closed
Native single-shot 30s video generation up to 4K, accepts up to 50 multimodal references; launched head-to-head vs MiniMax H3
Video Gen
2026-07-31
—
—
—
MiniMax
H3 (Hailuo H3)
Closed
Omni-modal video generator (text/image/video/audio input), up to 15s native 2K video with native stereo audio; open weights planned within days
Video Gen
2026-07-31
—
—
—
DeepSeekPublic-beta API release; retrained V4-Flash beats DeepSeek's own larger V4-Pro-Preview on all 9 published agent/coding benchmarks
Text
34/100
2026-07-31
1M
$0.14/1M tokens
$0.28/1M tokens
SenseTimeLightweight 8B-MoT open unified multimodal model (NEO-Unify architecture); adds native 4K image generation and stronger instruction-following edits vs prior U1
Image Gen
2026-07-31
—
—
—
Huawei
openPangu-2.0-Pro
Open
505B-param MoE (18B active), open-sourced with weights, inference code & technical report; trained 100% on Huawei Ascend 910B NPUs
Text
—
2026-07-31
512K
—
—
Alibaba
Qwen3.7-Flash
Closed
Native vision-language Flash upgrade with stronger object recognition, spatial intelligence, and multimodal agent execution
Multimodal
—
2026-07-25
1M
$0.03/1M tokens
$0.13/1M tokens
Ant Group
Ling-3.0-Flash
Open
124B-param MoE, only 5.1B active/token; hybrid-reasoning combines Ling speed + Ring reasoning; outperforms Ant's own 1T Ling-2.6 on 11/12 benchmarks; API-only (no open weights at launch), free on OpenRouter through 2026-08-03
Text
25/100
2026-07-23
256K
$0.021/1M tokens
$0.063/1M tokens
Alibaba
ABot suite
Closed
Amap's full-stack embodied AI upgrade: five models for navigation, manipulation, reasoning, memory, and motion control; claims SOTA on 17 benchmarks
Embodied
2026-07-22
—
—
—
Tencent
Hy-Embodied-VLA-0.5
Open
Vision-language-action model trained on 10,000+ hours of data, deployable across different robot platforms
Embodied
2026-07-19
—
—
—
Tencent
Hy-Embodied-VLM-1.0
Open
Perception model for robots; best-in-class on 19 of 38 benchmarks at 1/10th the compute of Tencent's prior flagship
Embodied
2026-07-19
—
—
—
TencentReasoning engine for embodied cognition: joint language-visual reasoning, world-state prediction, subgoal planning; ships with open RxBrain-Bench
Embodied
2026-07-19
—
—
—
SenseTime
SenseNova-U1 Pro
Closed
Unifies seeing, generating, acting via NEO-unify architecture; native 8K output; debuted at WAIC 2026
Multimodal
—
2026-07-19
—
—
—
Kunlun Tech
Mureka v9.5
Closed
AI music model with O3 reflective-reasoning + MuCo creation agent; pitched as least AI-sounding music generator, unveiled at WAIC
Multimodal
—
2026-07-19
—
—
—
Kunlun Tech
Matrix-Game 3.5
Closed
Interactive world-model upgrade shown at WAIC; 5B model, real-time 720p-class generation at 20FPS with 1-minute memory
Video Gen
2026-07-19
—
—
—
ModelBest
MiniCPM-RobotTrack
Open
0.9B on-device VLA model for person-tracking, unveiled at WAIC 2026; SOTA on 3 tracking benchmarks, runs 5+ FPS on Unitree Go2 onboard compute
Embodied
2026-07-17
—
—
—
ModelBest
MiniCPM-RobotManip
Open
1.5B on-device VLA model for robot manipulation, unveiled at WAIC 2026; outperforms larger models like π0.5 and Qwen-VLA per ModelBest
Embodied
2026-07-17
—
—
—
WeRide
WITT
Closed
Physical AI foundation model built on Atomic Physical Facts; claims 98% lower token cost and 200x data efficiency vs general models (self-reported, unverified)
Embodied
2026-07-17
—
—
—
Unisound
U2-Med
Closed
First tri-medical LLM spanning healthcare, insurance and pharma; built on U2 foundation model
Multimodal
—
2026-07-17
—
—
—
Moonshot AI
Kimi K3
Open
Flagship model, 2.8T total parameters, strong coding and agent capabilities, open weights by July 27
Multimodal
44/100
2026-07-16
1M
$3.00/1M tokens
$15.00/1M tokens
Ant Group
LingBot-VA 2.0
Open
Industry-first embodied-native causal video-action model; dual-stream mixture-of-transformers architecture, unifies frame prediction and policy execution from scratch
Embodied
2026-07-11
—
—
—
Ant Group
LingBot-World 2.0
Open
Open-source world model generating hour-long, no-decay 720p/60fps interactive experiences with built-in dynamic agents; up from minutes-level in v1.0
Embodied
2026-07-09
—
—
—
Alibaba
RynnWorld-4D
Open
DAMO Academy release; generates robot's predicted future as joint RGB + depth + optical-flow stream for manipulation planning, not just 2D video
Embodied
2026-07-08
—
—
—
Ant Group
LingBot-VLA 2.0
Open
6B open-source cross-embodiment model; trained on 60K hrs real-world data across 20 robot morphologies from 17 manufacturers
Embodied
2026-07-08
—
—
—
Ant Group
LingBot-Vision
Open
1.1B-param ViT-g/16 self-supervised vision model using masked boundary modelling, trained on ~161M curated images; 4 sizes released under Apache-2.0
Embodied
2026-07-07
—
—
—
Ant Group
LingBot-Depth 2.0
Closed
Next-gen monocular depth model paired with LingBot-Vision; claims to roughly halve robotic depth-estimation error in challenging scenarios
Embodied
2026-07-07
—
—
—
Tencent
Hy3 (Hunyuan 3.0)
Open
295B total / 21B active + 3.8B MTP layer, 192 routed experts (top-8) + 1 shared; official release upgrades Apr 2026 preview, deployed across WeChat, Yuanbao, WorkBuddy
Text
25/100
2026-07-07
256K
$0.18/1M tokens
$0.60/1M tokens
Shengshu Technology
Vidu S1
Closed
Real-time interactive video model with voice-controlled avatars; 540p up to 42fps; runs on consumer-grade GPUs instead of server clusters
Video Gen
2026-07-03
—
—
—
Zhipu AI (Z.ai)
GLM-5.2
Open
753B params; Rolled out to all GLM Coding Plan tiers (Lite/Pro/Max/Team); standalone API + MIT open weights
Text
51/100
2026-06-01
1M
$1.40/1M tokens
$4.40/1M tokens
MiniMax
MiniMax-M3
Open
428B params (23B active); 1-million-token context window; novel attention mechanism called MSA; native multimodality
Multimodal
29/100
2026-06
1M
$0.30/1M tokens
$1.20/1M tokens
Moonshot AI
Kimi-K2.7-Code
Open
1T total params / 32B active, 384 experts, ~30% lower reasoning-token usage vs K2.6
Code
26/100
2026-06
256K
$0.95/1M tokens
$4.00/1M tokens
StepFun
Step 3.7 Flash
Open
198B total / 11B active params; 1.8B ViT vision encoder; 3 reasoning levels (low/med/high); ~400 TPS; Apache 2.0; supports NVIDIA NIM on-prem deployment
Multimodal
—
2026-06
256K
$0.2/1M tokens
$1.15/1M tokens
Meituan
LongCat-2.0
Open
1.6T para + MoE; trained on 50,000 domestic Chinese chips; next-gen flagship
Text
19/100
2026-06
1M
$0.75/1M tokens
$2.95/1M tokens
XPeng
VLA 2.0
Closed
Second-generation vision-language-action world model — deliberative reasoning, controllable generation, and long-horizon forecasting. Runs on 4 self-developed Turing AI chips (~3,000 TOPS effective) in the Robotaxi platform.
Embodied
2026-06
—
—
—
Alibaba
Qwen3.7-Max
Closed
Flagship closed-source reasoning model. Ranked #1 on Artificial Analysis Intelligence Index (29/100) out of 218 models at release. 1M token context window. Text-only (no multimodal). Architecture reportedly dual-72B. Extended thinking / chain-of-thought mode. Released May 19, 2026.
Text
29/100
2026-05
1M
$2.50/1M tokens
$7.50/1M tokens
Baidu
ERNIE 5.1
Closed
MoE successor to ERNIE 5.0. ~⅓ total params, ~½ active params, 6% of pre-training cost. AIME 2026: 99.6% with tool use (#2 globally). Arena Search: 1,223 (#4 global, #1 China). Released May 9, 2026.
Text
—
2026-05
128K
$0.59/1M tokens
$2.65/1M tokens
Baichuan AI
Baichuan-M4
Closed
Medical AI flagship (May 2026). World #1 on HealthBench, HealthBench Hard & HealthBench Professional. Hallucination rate 3.3% — industry low via factuality-aware RL. Surpasses GPT-5.5, Opus 4.7, DeepSeek-V4-Pro. Powers 百小医 AI family doctor.
Text
—
2026-05
—
—
—
ModelBest
MiniCPM-V 4.6
Open
1.3B multimodal model (SigLip2 + Qwen3.5-0.8B). 260K token context. Runs on consumer phones (iOS/Android/HarmonyOS). Launched May 2026.
Multimodal
—
2026-05
260K
—
—
ModelBest
MiniCPM5-1B
Open
1B language model with 128K token context. SOTA on-device LLM at its size class. Apache 2.0.
Text
—
2026-05
128K
—
—
Zhipu AI (Z.ai)
GLM-5.1
Open
754B flagship — top scores on AIME 2026 (95.3) and GPQA-Diamond (86.2); strong agentic and coding ability
Text
—
2026-04
200K
$0.98/1M tokens
$3.08/1M tokens
MiniMax
MiniMax-M2.7
Open
229B MoE; self-evolving agent — 30% perf gain over 100+ rounds; agent teams, complex skills, 24/7 background agents
Agent
2026-04
1M tokens
$0.28/1M tokens
$1.20/1M tokens
Moonshot AI
Kimi K2.6
Open
1T MoE (32B active), 256K ctx; native multimodal agentic — top SWE-Bench (80.2) and AIME 2026 (96.4)
Multimodal
—
2026-04
256K
$0.60/1M tokens
$2.50/1M tokens
DeepSeek
DeepSeek-V4-Pro
Open
Flagship MoE: 1.6T total / 49B active params, 1M context, FP4+FP8 precision, built-in Think/Non-Think reasoning modes. Codeforces rating 3206.
Text
—
2026-04
1M
$0.435/1M tokens
$0.87/1M tokens
DeepSeek
DeepSeek-V4-Flash
Open
Efficient MoE: 284B total / 13B active params, 1M context. Same architecture as V4-Pro at far lower compute cost. Codeforces 2816.
Text
—
2026-04
1M
$0.14/1M tokens
$0.28/1M tokens
Baidu
ERNIE-Image
Open
8B single-stream Diffusion Transformer (DiT). Best text rendering in open source (LongTextBench: 0.9733). Up to 4K output. Apache 2.0. Released Apr 15, 2026.
Image Gen
2026-04
—
$0.03/image
—
Ant Group
Ring-2.6-1T
Open
1T total params, ~63B active per token; adaptive reasoning via high / xhigh modes; up to 66K output tokens; open weights (MIT); released May 8 2026; free tier available on OpenRouter
Text
—
2026-04
262K
$0.075/1M tokens
$0.625/1M tokens
Ant Group
Ling-2.6-1T
Open
1T total params, ~63B active per token; fast thinking approach cuts token cost to ~1/4 of comparable models; open weights (MIT); released Apr 23 2026; free tier available on OpenRouter via Novita AI
Text
17/100
2026-04
262K
$0.075/1M tokens
$0.625/1M tokens
Ant Group
Ling-2.6-flash
Open
Efficient sparse MoE: 104B total / 7.4B active, hybrid linear attention. 340 tokens/s on 4× H20. Strong agent and tool-use performance.
Text
—
2026-04
262K
$0.10/1M tokens
$0.30/1M tokens
Tencent
Hy3-preview
Closed
295B MoE (21B active); reasoning, coding, and agentic workloads. Via Tencent Cloud TokenHub. Released Apr 23, 2026.
Text
—
2026-04
256K
¥1.2/1M tokens
¥4.0/1M tokens
Xiaomi
MiMo-V2.5-Pro
Open
1.02T MoE, 42B active; KV-cache reduced ~7×; latest generation
Text
26/100
2026-04
1M
$0.0036/1M tokens
$0.087/1M tokens
AGIBot
Genie Operator-2 (GO-2)
Closed
AGIBot's new-generation embodied foundation model, powering autonomous manipulation and locomotion across its AGIBOT A2 and X2 robot lines.
Embodied
2026-04
—
—
—
Xiaomi
MiMo-V2-Pro
Closed
1T+ MoE, 42B active; hybrid attention; 1M token context
Text
—
2026-03
1M
$0.0036/1M tokens
$0.087/1M tokens
Unitree Robotics
UnifoLM-VLA-0
Open
Vision-Language-Action model enabling the G1 humanoid to autonomously perform household tasks from natural language commands. Runs onboard the robot. Open-sourced March 2026 — Unitree's first AI model release.
Embodied
2026-03
N/A
—
—
Kunlun Tech
SkyReels V4
Closed
Audio-visual creation model (Mar 2026). Dual-stream architecture; #1 globally on Text-to-Video (With Audio) and Image-to-Video (With Audio) tracks at release. Generates clips up to 3 minutes.
Video Gen
2026-03
—
—
—
Kunlun Tech
Mureka V9
Closed
Music generation model (Mar 2026). Paragraph-level text control; enhanced mixing quality, vocal expression, and style richness.
Audio
2026-03
—
—
—
Kunlun Tech
Matrix-Game 3.0
Closed
Physics-simulation interactive world model (Mar 2026). 5B params; 720P @ 40FPS; covers 1,000+ scenarios with Unreal Engine data. Industrial-grade real-time interactivity.
Text
—
2026-03
—
—
—
ByteDance
Seed2.0 Pro
Closed
ByteDance's flagship multimodal model. Understands text, image, and video. Ranks #3 globally on LMSYS Vision Arena and #6 on overall text arena. AIME 2025: 98.3, SWE-bench: 76.5%.
Multimodal
—
2026-02
272K
$0.47/1M tokens
$2.37/1M tokens
ByteDance
Seedream 5.0 Lite
Closed
Unified multimodal image generation with chain-of-thought visual reasoning, real-time web search, and native editing. Supports up to 14 reference images. Up to 4K output at 2048×2048.
Image Gen
2026-02
—
$0.026/image
—
ByteDance
Seedance 2.0
Closed
Unified audio-video joint generation model. Accepts text, image, video, and audio inputs; outputs native 2K video (up to 15s) with synchronized audio. 30% faster than Seedance 1.5 Pro.
Video Gen
2026-02
—
—
—
Alibaba
Qwen3.5
Open
Unified vision-language MoE family. Flagship: 35B-A3B (35B total / 3B active) and 397B-A17B. Thinking mode on by default. Supports image, video, text input. 201 languages. Apache 2.0.
Multimodal
—
2026-02
262K (1M w/ YaRN)
—
—
Alibaba
Qwen3-Coder-Next
Open
80B total / 3B active MoE coding agent. Excels at long-horizon agentic tasks, tool use, and IDE integration (Claude Code, Cline, Qwen Code). No thinking mode.
Code
—
2026-02
256K
—
—
Zhipu AI (Z.ai)
GLM-5
Open
754B MoE (40B active); 28.5T token pre-training; top-tier SWE-bench score of 77.8
Text
—
2026-02
200K
$1.40/1M tokens
$4.40/1M tokens
MiniMax
MiniMax-M2.5
Open
229B; SOTA SWE-Bench (80.2); 80% of MiniMax's own code generated by this model; M2.5-Lightning at 100 tok/s
Agent
2026-02
1M tokens
$0.30–$1.00/hr
$0.30–$1.00/hr
Baidu
ERNIE 5.0
Closed
2.4 trillion parameter unified multimodal (text + image + video + audio) in a single autoregressive framework. LMArena Text: 1,460 (#1 China, #8 global); Vision: 1,226 (#1 China, #8 global). Released Feb 6, 2026.
Multimodal
—
2026-02
128K
—
—
Ant Group
LingBot-VLA
Open
Open-source VLA foundation model trained on ~20,000 hours of real-world dual-arm robot data across 9 embodiments
Embodied
2026-02
—
—
—
Ant Group
Ring-2.5-1T
Open
World's first open-source 1T-parameter thinking model. Gold medal level at IMO 2025 (35/42 pts) and CMO 2025 (105/126). 3× throughput for sequences >32K.
Text
—
2026-02
256K
—
—
Ant Group
Ming-flash-omni-2.0
Open
Any-to-any omni model: accepts image, text, video, and audio; outputs image, text, and audio. 100B total / 6B active MoE. Supports zero-shot voice cloning and image generation/editing.
Multimodal
—
2026-02
—
—
—
Ant Group
LLaDA-2.1-flash
Open
Novel diffusion-based language model — not autoregressive. 103B params. Generates text by iterative token editing rather than left-to-right decoding. 102K monthly downloads.
Multimodal
—
2026-02
—
—
—
Baichuan AI
Baichuan-M3
Open
235B medical model (Feb 2026) built on Qwen3 architecture. Former world #1 on HealthBench. Hallucination rate 3.5%. Outperforms human doctors in diagnostic accuracy. 48GB VRAM with W4 quantization.
Text
—
2026-02
—
—
—
JD.com
JoyAI-LLM-Flash
Open
Open-sourced lightweight model — 4.8B total / 3B active params, pretrained on 20T text tokens. Released on HuggingFace as part of JD's open-source push.
Text
—
2026-02
—
—
—
Moonshot AI
Kimi K2.5
Open
1T MoE (32B active); trained on 15T vision+text tokens; thinking & instant modes; agent swarm support
Multimodal
—
2026-01
256K
$0.40/1M tokens
$1.90/1M tokens
DeepSeek
DeepSeek-OCR-2
Open
3B visual OCR model with document-to-markdown conversion, layout-aware grounding, and dynamic resolution. 1.66M monthly downloads.
Multimodal
—
2026-01
—
—
—
Tencent
HunyuanImage 3.0
Open
80B MoE (13B active); text-to-image and image-to-image with reasoning. Open weight on HuggingFace. Released Jan 26, 2026.
Image Gen
2026-01
—
—
—
Meituan
LongCat-Flash-Lite
Open
68.5B MoE, 2.9–4.5B active; fast and efficient
Text
—
2026-01
256K
—
—
AIsphere
PixVerse R1
Closed
Real-time world model; 1080p, <15s latency, physics-aware, infinite temporal continuity
Video Gen
2026-01
N/A
—
—
SenseTime
SenseNova V6.5 Omni
Closed
Real-time multimodal streaming model. Powers SenseChat's live audio/video interaction. Successor to V6 Pro.
Multimodal
—
2026-01
—
—
—
LimX Dynamics
FluxVLA Engine
Closed
LimX's Vision-Language-Action embodied-AI engine powering autonomous manipulation and locomotion on its humanoid and legged robots.
Embodied
2026
—
—
—
DeepSeek
DeepSeek-V3.2
Open
685B MoE with DeepSeek Sparse Attention (DSA). V3.2-Speciale variant won gold at 2025 IMO and IOI. 4.16M monthly downloads on HuggingFace.
Text
—
2025-12
128K
$0.252/1M tokens
$0.378/1M tokens
Kuaishou
Kling I2V 2.0
Closed
Image-to-video with industry-leading subject preservation
Multimodal
—
2025-12
N/A
$0.16/video (5s)
—
Xiaomi
MiMo-V2-Flash
Open
309B MoE, 15B active; trained on 27T tokens; open-source release
Text
22/100
2025-12
N/A
$0.01/1M tokens
$0.30/1M tokens
AIsphere
PixVerse V5.5
Closed
Text/image-to-video; HD output, multiple aspect ratios, character consistency
Video Gen
2025-12
N/A
—
—
Tencent
Hunyuan3D 3.0
Closed
3D asset generation from text, image, or sketch input
3D Gen
2025-11
—
—
—
iFlyTek
Spark X1.5
Closed
MoE reasoning model (29.3B total / 3B active parameters). Supports 130+ languages. Runs on a single Huawei Ascend server. Launched Nov 2025.
Text
—
2025-11
—
—
—
MiniMax
MiniMax-M2
Open
230B MoE (10B active); interleaved thinking; open-source under Modified MIT; free API available
Agent
2025-10
1M tokens
—
—
Infinigence AI
Megrez2-3x7B-A3B
Open
MoE edge model: 3×7B experts, 3B active parameters. Open-weight, Apache 2.0.
Text
—
2025-09
—
—
—
Alibaba
Qwen-Image
Open
Image generation and editing foundation model. Exceptional Chinese + English text rendering. Supports style transfer, object insertion/removal, layered editing, and depth/edge estimation. Released Aug 2025.
Image Gen
2025-08
—
—
—
ModelBest
MiniCPM-V 4.5
Open
8B multimodal model (Qwen3-8B + SigLIP2). Apache 2.0.
Multimodal
—
2025-08
—
—
—
Moonshot AI
Kimi k2
Open
Agentic model with tool use, web browsing, and code execution
Text
—
2025-07
128K
$0.57/1M tokens
$2.30/1M tokens
MiniMax
MiniMax-M1
Open
First open-weight large-scale hybrid-attention reasoning model; test-time compute scaling
Text
—
2025-06
1M tokens
$0.80/1M tokens
$2.20/1M tokens
MiniMax
Hailuo-02
Open
Physics-aware video generation; top-3 globally on VBench
Video Gen
2025-06
N/A
$0.06/video (5s)
—
Shengshu Technology
Vidu Q3
Closed
World's first storytelling-focused video model. Up to 16 sec, 1080p 24fps, native audio sync. Multilingual. Pro and Turbo variants. MCP integration.
Video Gen
2025-06
—
~$0.07/sec
—
Huawei
Pangu Ultra MoE (5.5)
Closed
718B-parameter MoE model (256 experts), trained on the full-stack hardware and software of Huawei Cloud's AI Cloud Service (CloudMatrix 384 supernodes). Part of the Pangu Models 5.5 family, which per Huawei Cloud's own release delivers upgrades across five capabilities: natural language processing, computer vision, multimodal, prediction, and scientific computing.
Text
—
2025-06
—
—
—
RedNote (Xiaohongshu)
dots.llm1
Open
RedNote's first open-source LLM — 142B total / 14B active params (MoE), pretrained on 11.2T high-quality tokens with no synthetic data. Performance comparable to Alibaba's Qwen2.5-72B.
Text
—
2025-06
32K
—
—
DeepSeek
DeepSeek-R1-0528
Open
Latest R1 update: 685B, adds system prompt support, deeper reasoning (23K avg tokens on AIME vs 12K prior). Distilled 8B version achieves 86% on AIME 2024.
Text
—
2025-05
128K
$0.50/1M tokens
$2.15/1M tokens
Kuaishou
Kling 2.0
Closed
4K-capable video generation; advanced camera controls and scene coherence
Video Gen
2025-04
N/A
$0.16/video (5s)
—
Kuaishou
Kolors 2.0
Open
Upgraded photorealistic image model; open-source available
Image Gen
2025-04
N/A
$0.004/image
—
SenseTime
SenseNova V6 Pro
Closed
620B MoE hybrid. Real-time audio/video streaming. Ranked #1 in China in multimodal reasoning at launch. Lowest reasoning cost in industry at launch (Apr 2025).
Multimodal
—
2025-04
256K
¥2.8/1M tokens
¥8.4/1M tokens
Kunlun Tech
Skywork-OR1-32B
Open
Open-source math and coding reasoning model (Apr 2025). 32B params; rivals DeepSeek-R1 on competitive programming benchmarks. 7B variant also available.
Text
—
2025-04
—
—
—
Qihoo 360
360Zhinao3-7B
Open
Latest open-weight 360Zhinao generation (Base / Instruct / O1.5 long chain-of-thought reasoning variant). Free for commercial use.
Text
—
2025-04
32K
—
—
Baidu
ERNIE X1
Closed
Baidu's first dedicated reasoning model with extended chain-of-thought
Text
—
2025-03
128K
$0.28/1M tokens
$1.10/1M tokens
Baidu
ERNIE 4.5
Closed
Improved multimodal and reasoning; best ERNIE for Chinese enterprise tasks. The flagship ERNIE 4.5 model is closed-source, though several variants in the ERNIE 4.5 family are openly available on HuggingFace.
Multimodal
—
2025-03
128K
$0.55/1M tokens
$2.20/1M tokens
StepFun
Step-3
Closed
StepFun's 2025 flagship; 1M token context and native reasoning mode
Text
—
2025-01
1M
$0.57/1M tokens
$1.42/1M tokens
Baichuan AI
Baichuan-M2
Open
32B medical reasoning model (2025) built on Qwen2.5-32B with innovative Large Verifier System for real-world clinical reasoning.
Text
—
2025-01
—
—
—
ModelBest
MiniCPM-o 2.6
Open
8B any-to-any model (text + vision + speech). Surpasses GPT-4o and Gemini 1.5 Pro on single-image understanding benchmarks. Launched Jan 2025.
Multimodal
—
2025-01
—
—
—
Kuaishou
Kling 1.6 Pro
Closed
Previous flagship; supports 1080p, 5-10s clips, camera controls
Video Gen
2024-12
N/A
$0.14/video (5s)
—
Infinigence AI
Megrez-3B-Omni
Open
3B omni model processing text, vision, and audio. Outperforms LLaVA-NeXT-Yi-34B on vision benchmarks. Optimised for on-device and edge deployment.
Multimodal
—
2024-12
—
—
—
Kunlun Tech
Skywork-o1
Open
Reasoning model (Nov 2024) — among China's first o1-style models with chain-of-thought Chinese logical reasoning. Open-source 8B variant (Llama 3.1 base) plus proprietary advanced version.
Text
—
2024-11
—
—
—
Qihoo 360
360Zhinao2-7B
Open
Second-generation 360Zhinao with Base and Chat variants at 4K / 32K / 360K context lengths.
Text
—
2024-11
360K
—
—
Baichuan AI
Baichuan4-Turbo
Closed
Enterprise flagship general LLM (late 2024). 10%+ usability gain vs prior generation; priced at ~80% of GPT-4o. Supports multimodal input.
Text
—
2024-10
192K
—
—
Baichuan AI
Baichuan4-Air
Closed
MoE variant (PRI architecture) of Baichuan4. High performance at low cost for API deployments.
Text
—
2024-10
192K
—
—
01.AI
Yi-Lightning
Closed
MoE flagship API model (Oct 2024). Ranked #6 on Chatbot Arena at launch — joint 3rd among LLM companies. 40% faster inference than prior Yi models. Final model before 01.AI halted pre-training in early 2025.
Text
—
2024-10
16K
$0.14/1M tokens
$0.14/1M tokens
01.AI
Yi-Coder
Open
Code model (Sep 2024); 1.5B and 9B variants; supports 52 programming languages; 128K context window.
Code
—
2024-09
128K
—
—
iFlyTek
Spark 4.0
Closed
Flagship LLM (Aug 2024). Claims comparable performance to GPT-4 Turbo on Chinese language benchmarks.
Text
—
2024-08
—
—
—
StepFun
Step-2
Closed
Very long context window; strong document understanding
Text
—
2024-07
256K
¥0.038/1K tokens
¥0.12/1K tokens
StepFun
Step-1X-Image
Closed
High-resolution image generation model
Image Gen
2024-07
N/A
¥0.04/image
—
01.AI
Yi-1.5
Open
Open-source series (May 2024); 6B–34B variants. Improved coding, math, reasoning, and instruction-following over original Yi.
Text
—
2024-05
—
—
—
Qihoo 360
360Zhinao-Search
Open
Retrieval + 1.8B reranking models that ranked #1 in the Retrieval and Reranking tasks of the C-MTEB leaderboard.
Embedding
2024-05
—
—
—
01.AI
Yi-VL-34B
Open
34B vision-language model (early 2024). Open-weight multimodal extension of Yi-34B.
Multimodal
—
2024-01
—
—
—
01.AI
Yi-34B
Open
Founding open-source model (Nov 2023). Topped HuggingFace Open LLM Leaderboard and C-Eval at launch. 200K-token context variant (Yi-34B-200K) also available.
Text
—
2023-11
200K
—
—
Kunlun Tech
Skywork-13B
Open
Founding open-source bilingual LLM (Oct 2023). 13B params; pretrained on 3.2T tokens. Led same-scale models on CEVAL, CMMLU, and MMLU at launch.
Text
—
2023-10
—
—
—
Baichuan AI
Baichuan2
Open
Open-source bilingual LLM series (Sep 2023); 7B and 13B variants; trained on 2.6T tokens. Available on Hugging Face under permissive license.
Text
—
2023-09
4K
—
—
JD.com
ChatRhino (言犀)
Closed
JD's 100B-parameter industry-focused LLM, blending 70% general data with 30% native supply-chain intelligence data for retail, logistics, finance, healthcare, and city-services applications.
Text
—
2023-07
32K
—
—
Meituan
LongCat-Image
Open
Image generation model, ~6B params; data-quality focused; released Dec 2025
Image Gen
—
N/A
—
—
MeituanChain-of-thought reasoning variant of LongCat-Flash
Text
—
—
256K
—
—
Meituan
LongCat-Flash-Chat
Open
560B MoE, 27B active; open-source; 500K free tokens/day
Text
—
—
128K
—
—
WeRide
WeRide GENESIS
Closed
Generative Engineered Neural Environment for Simulated Intelligence in Self-driving. WeRide's proprietary general-purpose simulation platform combining physical AI with generative AI. Rapidly builds photorealistic simulated cities in minutes, generates diverse edge-case scenarios from billions of km of real-world data, and models realistic pedestrian and driver behavior — all at centimeter-level fidelity. Supports L2++ through L4 AV development and validation via four modules: AI Scenarios, AI Agents, AI Metrics, and AI Diagnosis.
Embodied
—
—
—
—
Pony.ai
PonyWorld 2.0
Closed
Second-generation world model underpinning the Virtual Driver L4 autonomous driving platform. Introduces an Intention layer — a structured representation of decision-making that enables the system to evaluate its own driving decisions, identify accuracy gaps across scenarios, and direct targeted data collection rather than broad undirected improvement. Direct sensor-to-action architecture with no language models in the inference pipeline; runs on 1016 TOPS across three NVIDIA DRIVE Orin-X SoCs with redundant failover.
Embodied
—
—
—
—
Unisound
UniGPT (Shanhai 山海)
Closed
60B+ parameter general large model (v5.0); underpins all Unisound vertical products. Medical, enterprise, and consumer deployments.
Text
—
—
—
—
—
Unisound
U2-ASR 2.5
Closed
First LLM-based semantic ASR model for Chinese. Covers 100+ dialects across 7 dialect systems. >90% accuracy. Available via Token Hub API.
Audio
—
—
—
—
Unisound
U2-TTS / U2-TTS-Clone
Closed
Text-to-speech and voice cloning with full-duplex millisecond response. Available via Token Hub API.
Audio
—
—
—
—
Unisound
U1-OCR
Closed
Industrial-grade document intelligence model for OCR and document understanding. Launched February 2026.
Multimodal
—
—
—
—
—
Disclosed, 2026 YTD
Funding & Markets
Private funding rounds and public market data for China's leading AI companies
43
Companies tracked
23 public · 20 private
23 public · 20 private
~$34.9B
AI startup funding
12 AI-native cos
12 AI-native cos
Compare companies →
Valuation, total funding & AI ARR, side-by-side
Most Active Investors
1
Tencent
16
2
HongShan Capital
12
3
Alibaba
12
4
IDG Capital
9
5
Shunwei Capital
7
Derived live from 170 rounds
ByteDance
PRIVATE
Valuation$600B
Total Raised~$42.21B+
AI ARR$4B
Alibaba
BABA / 9988.HK
Market Cap—
Cloud Revenue (FY25)¥158B
AI ARR¥10B
Zhipu AI (Z.ai)
2513.HK
Market Cap—
Total Raised$10.79B
AI ARR$1.6B
MiniMax
0100.HK
Market Cap—
Total Raised$2.17B
AI ARR$800M
Moonshot AI
PRIVATE
Valuation$35B
Total Raised~$8.2B+
AI ARR$1B
DeepSeek
PRIVATE
Valuation$50B
Total Raised$7.35B
AI ARR$1B
StepFun
PRIVATE
Valuation$10B
Total Raised$3.23B+
Last Round2026
Baidu
BIDU / 9888.HK
Market Cap—
ExchangeNASDAQ: BIDU · HKEX: 9888
Founded2000
Ant Group
PRIVATE
Valuation$79B
Total Raised$20.53B+
Last Round2020
Tencent
0700.HK
Market Cap—
Revenue (FY25)¥751.8B
Founded1998
Kuaishou
1024.HK
Market Cap—
ExchangeHKEX: 1024
Kling ARR$500M
Xiaomi
1810.HK
Market Cap—
Revenue (FY25)¥457.3B
Founded2010
Meituan
3690.HK
Market Cap—
Revenue (FY25)¥364.9B
Founded2010
AIsphere
PRIVATE
Valuation$2B
Total Raised~$669M+
Last Round2026
Unitree Robotics
688836.SS
Market Cap—
FY2025 Revenue¥1.71B
Founded2016
SenseTime
0020.HK
Market Cap—
Revenue (FY25)¥5.01B
Founded2014
Baichuan AI
PRIVATE
Valuation$2.7B
Total Raised$1.04B
Last Round2024
01.AI
PRIVATE
Valuation$1B
Total Raised$300M+
Last Round2024
Kunlun Tech
300418.SZ
Market Cap—
Revenue (FY25)¥8.2B
Founded2008
Evoken
PRIVATE
Valuation$2B
Total Raised$430M+
Last Round2026
iFlyTek
002230.SZ
Market Cap~¥190B
FY2025 Revenue¥27.1B
Founded1999
Infinigence AI
PRIVATE
ValuationUndisclosed
Total Raised~$251M+
Last Round2026
Enflame Technology
688801.SS
Market Cap—
FY2025 Revenue¥990.2M
Founded2018
SiliconFlow
PRIVATE
Valuation¥7.7B
Total Raised$419M
Revenue¥55.3M
Cambricon
688256.SS
Market Cap—
FY2025 Revenue¥6.5B
Founded2016
ModelBest
PRIVATE
Valuation¥20B
Total RaisedUndisclosed
Last Round2026
Moore Threads
— (SSE)
Market Cap—
FY2025 Revenue¥1.505B
Founded2020
WeRide
WRD / 0800.HK
Market Cap—
FY2025 Revenue~$93.6M
Founded2017
Pony.ai
PONY / 2026.HK
Market Cap—
FY2025 Revenue$90.0M
Founded2016
Shengshu Technology
PRIVATE
ValuationUndisclosed
Total Raised~$440M
Last Round2026
DiDi Autonomous Driving
PRIVATE
ValuationUndisclosed
Total Raised$1.39B+
Last Round2025
Unisound
9678.HK
Market Cap—
FY2025 Revenue¥1.21B
Founded2012
Momenta
06880.HK
Market Cap—
FY2025 Revenue¥2.4B
Founded2016
Qihoo 360
601360.SS
Market Cap~$10.1B
Market Cap~$10.1B
Founded2005
CXMT
688825.SS
Market Cap—
IPO Proceeds~$8.6B
Founded2016
Manus (Butterfly Effect)
PRIVATE
Valuation~$500M
Total Raised$75M
Last Round2025
LimX Dynamics
PRIVATE
Valuation¥15B
Total Raised~$400M+
Last Round2026
XPeng
XPEV / 9868.HK
Market Cap—
ExchangeNYSE: XPEV · HKEX: 9868
Founded2014
AGIBot
PRIVATE
Valuation~$1.5B+
Total Raised$84M+
Last Round2025
Huawei
PRIVATE
FY2025 Revenue¥880.9B
R&D Investment¥192.3B
Founded1987
Pragmatik Labs
PRIVATE
Valuation$2B
Total Raised$220M
Last Round2026
RedNote (Xiaohongshu)
PRIVATE
Valuation~$70B
Total Raised~$917M+
Last Round2026
JD.com
JD / 9618.HK
Market Cap—
FY2025 Revenue¥1,309.1B
Founded2004
Data methodology: Funding data on this page is compiled from publicly available sources — company press releases, regulatory filings, and media reports — and fact-checked by Tony Peng with the assistance of AI. Where a round appears in both a company disclosure and press report, company disclosure takes precedence. All figures may be approximate.
Shipped Products
Applications
AI products from China's leading companies, by category
Est. 2026
The essential tracker for China's AI ecosystem
One place to follow every major model release, funding round, product launch, and breakthrough from China's AI ecosystem.
Tony Peng
AI observer · Journalist · Communications professional
China AI Index is a research project built and maintained by Tony Peng — a long-time observer of the AI industry with a background in journalism and communications — with the assistance of AI (Claude Code, Qoder, and Z Code). Tony is also the writer behind Recode China AI, a newsletter tracking the latest advances and news stories in intelligent technologies.
It exists to give English-speaking researchers, investors, journalists, and developers a structured reference for China's AI ecosystem — most authoritative information is published in Chinese, and this site translates and synthesizes it.
What we track
Models
Every major LLM, reasoning model, vision model, and generative AI model from Chinese labs — with benchmark scores, context windows, and live API pricing so you can compare apples to apples.
Financing
Full funding histories for private companies (ByteDance, DeepSeek, MiniMax, Moonshot…) and market data for public ones (Alibaba, Baidu, Tencent, Kuaishou). Valuations, investors, round sizes.
Newsfeed
News covering model launches, research breakthroughs, product announcements, and strategic moves — presented in a clean timeline with company tags and category filters.
People
Key founders, CEOs, CTOs, and researchers behind China's AI labs — with titles, Chinese names, and links to their social profiles so you know the faces shaping the industry.
Members
Regulation
A timeline of China's AI rules — generative AI measures, algorithm filing requirements, model approval lists, data law touchpoints, and provincial subsidy programs. Click any entry for the original text.
Members
Chips
Domestic AI accelerators, training clusters, and memory chips — specs, interconnect bandwidth, process nodes, and deployment status for China's emerging chip ecosystem.
Data Disclaimer
All pricing, benchmark scores, funding amounts, and valuations are sourced from public company announcements, tech media reporting, and developer documentation. They are provided for informational purposes only and may be approximate or subject to change.
Stock prices shown are approximate historical ranges — not real-time quotes. Always verify with a live financial data source before making investment decisions. Benchmark scores reflect results at the time of model release and may not account for subsequent updates.
This database is independently maintained and has no affiliation with any of the companies listed. Information is updated manually on a best-effort basis.
China AI Regulation
A timeline of generative AI rules, algorithm filing requirements, model approval lists, and data law touchpoints — with links to original texts
2026
Sep
Extends the August deployment-side guide upstream to agent system development. Together the two documents suggest TC260 is building a build-and-run pair for agents ahead of any binding rule.
View Official Text ↗
2026
Sep
Four documents released together, the start of a general-plus-sector series. The General Principles apply across all industries; the other three give sector-specific security practice guidance for education, health, and broadcasting/online audiovisual. Drafting units include CESI, the NRTA research academy, NHC and MOE information centers.
View Official Text ↗🔒 14 more regulations — Paid Subscribers Only
You're viewing the first 2 entries free. The full regulation timeline — 16 laws, measures, and strategy documents back to 2017 — is included with a paid Recode China AI subscription.
China's Chip Ecosystem
Domestic AI accelerators, large-scale training clusters, and memory chips — tracking the hardware layer behind China's AI buildout.
AI Accelerators
Clusters & Supernodes
Memory & HBM
Role
Precision
| Company | Chip | Role | Peak Compute | Memory Tech | Capacity | Bandwidth | Interconnect Bandwidth |
TDP | Process | Status | Highlights |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Huawei ↗ | Ascend 910B |
Training |
256 TFLOPS
BF16
|
HBM2e | 64 GB | 2.0 TB/s |
392 GB/s
HCCS 3.0
|
400 W | SMIC N+2 | Production | Most-deployed domestic training chip. Used at Alibaba, ByteDance, Baidu, and Tencent. |
| Huawei ↗ | Ascend 910C |
Training |
800 TFLOPS
FP16
|
HBM2e | 96 GB | 4.0 TB/s |
800 GB/s
HCCS
|
900 W | SMIC N+2 | Production | Dual-die upgrade to 910B. Powers the Atlas 900 A3 SuperPoD (300 PFLOPS). |
| 🔒 18 more accelerators — paid subscribers only | |||||||||||
Specs marked ~ are estimates from third-party analysis. Click any column header to sort. Interconnect Bandwidth = chip-to-chip fabric bandwidth per chip (bidirectional).
| Cluster / Supernode | Operator | Chip | Chip Count | Total Compute | Interconnect Fabric | Notes |
|---|---|---|---|---|---|---|
Atlas 900 A3 SuperPoD (CloudMatrix 384) |
Huawei Cloud | Ascend 910C | 384 |
300 PFLOPS
FP16
|
All-optical (OXC) | Huawei's current-generation supernode. Uses optical circuit switching (OXC) for low-latency inter-chip communication. Announced April 2025 as a full-stack domestic alternative to Nvidia GB200 NVL72. |
Atlas 950 SuperPoD |
Huawei Cloud | Ascend 950 | 8,192 |
8 EFLOPS
FP8
|
16.3 PB/s aggregate | Next-generation supernode targeting 2026 delivery. 1,152 TB total memory. Full liquid cooling. Footprint larger than two basketball courts. Would represent a major leap in domestic training capacity if delivered at spec. |
| 🔒 3 more clusters — paid subscribers only | ||||||
Cluster sizes and compute figures are from official announcements or credible press reports.
| Company | Product | Type | Capacity | Bandwidth | Process | Status | Notes |
|---|---|---|---|---|---|---|---|
| CXMT ↗ | DDR5 |
DDR | 16–32 Gb dies | ~6400 MT/s | ~16nm (G4) | Production | CXMT's latest-generation DDR5, unveiled Nov 2025 — its push into the high-end DRAM market against Samsung, SK Hynix, and Micron. |
| CXMT ↗ | LPDDR5X |
LPDDR | Mobile | ~8533 MT/s | ~16nm | Production | Low-power mobile DRAM for smartphones and edge/AI devices. |
| 🔒 2 more memory chips — paid subscribers only | |||||||
DRAM, HBM, and NAND from Chinese memory makers. Specs marked ~ are estimates from third-party analysis or press reports.