DeepSeek news
DeepSeek is a Chinese AI lab known for strong, low-cost open models. All our DeepSeek stories, newest first.
- Nvidia
- OpenAI and ChatGPT
- Anthropic and Claude
- Google Gemini
- Meta AI and Llama
- Qwen and Alibaba AI
- Microsoft and Copilot
- AI agents
- AI chips
- DeepSeek
- Mistral AI
- xAI and Grok
- Apple AI
- Amazon and AWS AI
- Open-source AI
- AI coding tools
- AI funding and deals
- AI policy and rules
- AI research
- Learn AI
All DeepSeek stories
- DeepSeek alum launches open-source personal AI agent for local PCs
Linger allows users to run high-parameter models on consumer hardware while keeping all personal data on local devices.
- DeepSeek closes 100 billion yuan funding round with 31 investors
The massive funding round includes major industrial players like Tencent and CATL, signaling a push for industrial synergy in China's AI landscape.
- DeepSeek's 100-billion-yuan funding round closes, with 31 outside investors named
Corporate records show DeepSeek's second big round has landed, backed by telecom giants, a battery maker, a phone maker and investment funds. Here is why it matters.
- DeepSeek removes separate fees for model-triggered web search
The AI company is eliminating extra costs for search and web fetching when triggered by its models, though standalone API calls remain paid.
- DeepSeek's free chat now cuts off long conversations left idle for 7 days
DeepSeek's free app and web chat now limits old, long conversations, and users want the 7-day rule dropped. It is a reminder not to treat one chat thread as storage.
- A free browser extension wires web chatbots to tools on your own computer
Web MCP Bridge lets free chat tabs like DeepSeek, Gemini, ChatGPT and Claude use local tools, with no API keys. It is clever, but it is for tinkerers.
- Alibaba's QwenWork lets users pick rival Chinese AI models for office tasks
Alibaba's office agent now offers models from Kimi, GLM, DeepSeek, MiniMax and Step, and companies can set who sees which ones.
- Alibaba's QwenWork lets users pick rival Chinese AI models for office tasks
Alibaba's office agent now offers models from Kimi, GLM, DeepSeek, MiniMax and Step, and companies can set who sees which ones.
- DeepSeek blames low computing power for chat limits, raising doubts about free service
DeepSeek is now openly citing a lack of computing power to cap long chats, which puts its free model's future in question.
- DeepSeek caps long chats and daily use in its app and website, citing too little computing power
DeepSeek has reportedly limited long conversations and daily use in its app and web chat, so anyone who pastes long documents should adjust now.
- DeepSeek's chat site code reveals a full usage-quota system, and hints at a paid plan
Developers found quota-tracking code in DeepSeek's chat front end. It shows how limits will appear to users, but DeepSeek has announced no pricing.
- DeepSeek users report new caps on long chats and daily use
Users say DeepSeek's app and website now limit long conversations, input size and daily use, a sign that demand is outrunning computing capacity.
- DeepSeek caps chat history at 7 days and input at about 40,000 characters
DeepSeek's app and website now limit chat history, input length and daily use, reportedly because computing power is strained. Long-document habits may need to change.
- Qwen Office adds 10 models, including rivals' Kimi, GLM and DeepSeek
Alibaba's office AI tool now lets users pick the model for each task, so a hard job and a quick one need not run on the same engine.
- DeepSeek caps prompt length and old chats, blaming a shortage of computing power
DeepSeek's app and web chatbot now limits long prompts and stale chats because it lacks computing power, which affects anyone who works on long projects.
- DeepSeek adds search, OCR and speech tools to its API
The Chinese AI provider launched a suite of integrated tools for developers, but the move introduces specific technical limits and data privacy considerations.
- dsh-purge adds local controls for DeepSeek Harness plugins
A new community add-on provides a dashboard and command-line tools to manage, backup, and roll back configurations for self-hosted DeepSeek Harness setups.
- DeepSeek lets developers plug its search and file-reading features into their own apps
DeepSeek's app features, like web search and file reading, can now be called by other software, with extra charges on top of normal model fees.
- DeepSeek web code reveals potential new subscription plan
Developers discovered hidden logic in DeepSeek's web interface suggesting a 'Coding Plan' despite no official announcement.
- Lawyer optimizes private AI server to run DeepSeek V4.1 Flash with higher fidelity
By upgrading hardware and increasing bit precision, a trial lawyer successfully improved the accuracy of a local AI model for legal workflows.
- DeepSeek staff rarely leave to found startups, and a report says why
While other Chinese AI labs see staff leave to found startups, DeepSeek's researchers mostly stay or join big firms. Pay, structure and investors explain it.
- Zhipu and DeepSeek are putting Chinese chips through the hardest tests
Two leading Chinese model labs are using home-grown chips for demanding work, which a commentary argues is what those chips have always lacked.
- DeepSeek's website code hints at its first paid chat plan, but nothing is confirmed
Developers found quota-limit logic in DeepSeek's chat front end and guess a paid consumer plan is coming. DeepSeek has said nothing, but the clue touches how it makes money.
- DeepSeek opens its app features to developers through its API
DeepSeek's API now handles web search, file reading and more on its own servers, so developers can build a close copy of its chat app without wiring those tools up themselves.
- DeepSeek puts its app's search, file and voice tools into its API and opens its chat frontend
Developers can now add web search, document reading and voice to their own apps with a few settings, but the tools only work with one model and the bills can surprise.
- A fan-made whale mascot is helping DeepSeek Harness win developers
DeepSeek says its Harness tool has 1.54 million daily users, and a community-made anime mascot and a plugin-heavy design look like part of the reason.
- DeepSeek opens full app toolkit to developers via API
The Hangzhou-based AI lab is allowing third-party builders to integrate web search, file parsing, and voice functions into their own products.
- DeepSeek investment links raise Nvidia security concerns
Allegations regarding investments from entities linked to Alice Huang suggest potential ways for Chinese AI firms to access advanced Nvidia chips.
- Firms tied to Alice Huang reportedly invested in DeepSeek to secure Nvidia chips
A report says a network linked to Huang Le joined DeepSeek's funding round, tying a Chinese AI lab to a known supplier of black-market Nvidia chips.
- Alice Huang's companies invest in DeepSeek and supply Nvidia chips
A Chinese businesswoman under investigation for alleged espionage is linking major AI firms and hardware giants.
- Firms linked to Alice Huang invest in DeepSeek to supply Nvidia chips
Companies connected to a figure linked to Nvidia's former CEO reportedly invested in DeepSeek to facilitate black market chip access.
- Alice Huang's firms invest in DeepSeek with promised Nvidia chips
A Singapore-based cloud firm controlled by a Chinese businesswoman is under scrutiny for facilitating billion-dollar GPU deals for Chinese AI labs.
- Alice Huang's business network linked to restricted Nvidia chip supply
Reports suggest companies tied to Alice Huang may be facilitating the movement of restricted Nvidia Blackwell chips to the Chinese AI startup DeepSeek.
- Nvidia partner under investigation over DeepSeek investments and chip flows
Alice Huang's Megaspeed is facing scrutiny as it links high-end Nvidia GPU access to investments in a leading Chinese AI lab.
- Doubao Work adds Kimi, GLM and DeepSeek models and prices each by a quota multiplier
ByteDance's Doubao Work now lets users pick third-party models, each drawing down allowance at its own rate, so model choice becomes a cost decision.
- Jailbroken DeepSeek AI agent evades antivirus in telecom attack
An AI agent bypassed security measures to reach a healthcare database by autonomously creating new tools and scripts when its original methods were blocked.
- Hackers used DeepSeek AI to steal 2,000 defense documents
A cyber espionage group called Silent Lynx breached Central Asian government systems using an AI agent to automate hacking tasks.
- DeepSeek V4.1 Pro is only a name, so a fair test pits MIDM K 3.0 Pro against V4 Pro
A router vendor says DeepSeek V4.1 Pro has no model card, price or endpoint, so comparisons with it are guesses. Here is what can actually be tested.
- Google's Android coding benchmark now lists Kimi K3.1 Pro and DeepSeek V4.1 Flash
Google's Android Bench now has scores for Kimi K3.1 Pro and DeepSeek V4.1 Flash, showing a clear trade-off between accuracy, speed and cost.
- CXMT chairman reportedly joined DeepSeek's second funding round
Digitimes reports that the head of a major Chinese memory chipmaker took part in DeepSeek's funding, hinting at ties between chip money and AI labs.
- DeepSeek agent platform reaches 1.54 million daily users
- Jailbroken DeepSeek AI agent breaches European telecom and healthcare data
- Tencent's Workbuddy refunds half of DeepSeek-V4.1 Flash credits after a quiet promo change
- Chinese AI chip stocks rise as DeepSeek releases software for domestic chips
- DeepSeek Harness plans an official plugin store as its self-reported user count climbs
- DeepSeek's open-source agent app Harness reaches 1.54 million daily users
- DeepSeek agent platform reaches 1.54 million daily active users
- Open-weight DeepSeek model nears OpenAI's cyber capabilities at lower cost
- Cheap AI models compared: DeepSeek Flash and Claude Haiku tie on accuracy, not on cost
- Zhu Yiming joins DeepSeek's second funding round, Securities Star reports
- DeepSeek appears to be quietly testing V4.1 Pro, and early benchmark results look strong
- Open-weight DeepSeek model nearly matches OpenAI's restricted cyber tool
- DeepSeek's ownership and the impact of its efficient AI models
- Open-weight DeepSeek model nearly matches closed AI in cyber tests
- DeepSeek aims to accelerate AGI development through open source models
- DeepSeek Harness reaches 1.5 million daily active users
- Fixstars offers DeepSeek's open model on GPU servers inside Japan
- DeepSeek Harness reaches 1.54 million daily users and plans a plugin marketplace
- DeepSeek V5 release remains unconfirmed with 2027 target
- Fans built a 'Big Fat Fish' mascot for DeepSeek, and it is flooding Chinese social media
- DeepSeek reportedly plans to train its next model on Huawei Ascend chips
- DeepSeek engineer says open source is a race against time, and admits it adds safety risk
- Chery launches Jetour T1 SUV with DeepSeek AI and plug-in hybrid options
- DeepSeek argues open source is the fastest path to AGI
- DeepSeek open-sources DeepInfer to accelerate AI development
- DeepSeek's Super-agent design lets AI sort and start tasks without human prompts
- DeepSeek outlines a 'Super Agent' that would direct other AI agents with fewer human check-ins
- DeepSeek reportedly plans to train its V5 model on Huawei chips instead of NVIDIA's
- China-linked hackers use DeepSeek AI to hit 200 organizations in 14 countries
- Chinese AI leaders debate open-source risks and benefits at Wuzhen Summit
- DeepSeek releases open-source AI agent to automate coding tasks
- China to accelerate comprehensive AI law and safety standards
- DeepSeek plans to pre-train next-generation model on Huawei Ascend chips
- DeepSeek Harness adds a creation mode for building AI plugins from a sentence
- DeepSeek releases open-source Super-agent for autonomous software tasks
- DeepSeek releases open-source Super-Agent for autonomous coding
- DeepSeek open-sources Super-Agent prototype for autonomous coding
- DeepSeek releases dsh-superagent to manage complex coding tasks
- DeepSeek launches Super-Agent to automate software engineering
- OpenAI adjusts funding goals as rivals move in on market share
- DeepSeek open-sources Super-Agent to handle GitHub tasks via chat
- Which AI chatbot is best? A Singapore writer tests four and finds it depends on you
- A hands-on test of DeepSeek's Super-agent: it fixed a bug nobody asked it to fix
- DeepSeek releases open-source coding agent that fixes bugs without being asked
- DeepSeek debuts Super-Agent to compete with Anthropic's Claude Code
- DeepSeek proposes super agent to coordinate multiple AI assistants
- DeepSeek open-sources dsh-superagent research prototype
- DeepSeek launches open-source Super-Agent for autonomous coding
- DeepSeek proposes blueprint to manage multiple AI coding agents
- DeepSeek open-sources Super-Agent for team-level coding
- DeepSeek prototype Super-Agent proactively suggests code fixes
- DeepSeek launches autonomous coding agent to compete with Anthropic
- DeepSeek open-sources Super-Agent to automate team coding tasks
- DeepSeek launches Super Agent to manage multiple AI tasks simultaneously
- DeepSeek launches Super Agent to challenge Anthropic's coding tools
- DeepSeek releases dsh-superagent to manage coding tasks in parallel
- DeepSeek launches autonomous coding agent to challenge Anthropic
- DeepSeek Harness releases research prototype of autonomous coding super-agent
- DeepSeek shows an AI manager that hands work to AI workers and fixes bugs unprompted
- DeepSeek outlines Super-agent, an AI helper meant to work across chat and code on its own
- DeepSeek outlines Super-agent, a design where one AI coordinates other AI workers
- DeepSeek's open-source release hands out the machinery that runs AI, not just the model
- DeepSeek describes a Super-agent: one AI dispatches work, several others carry it out
- DeepSeek publishes the design for an agent that works without being asked
- DeepSeek unveils Super-agent architecture for parallel task management
- DeepSeek open sources architecture for proactive coding agent
- DeepSeek releases open-source coding agent for parallel GitHub tasks
- DeepSeek and OpenAI launch new models with divergent pricing models
- Ask AI if Mao was a dictator and the answer depends on where the AI was built
- Hengdian Capital joins DeepSeek's second funding round, which China.com calls a record
- DeepSeek reportedly closes a second funding round at a 615.5 billion yuan valuation
- DeepSeek V4.1 Flash can run at home on 64GB, but speed is the open question
- DeepSeek raises $16 billion in funding at $74 billion valuation
- U.S. lawmakers propose ban on DeepSeek AI for government devices
- DeepSeek Harness adds plugin for delegating tasks to Claude and Codex
- DeepSeek's V4.1 Pro has a name but no model card, price or release date
- DeepSeek's cache discount is a bigger cost lever than its peak-hour surcharge
- 01.AI takes state funding and moves its headquarters to DeepSeek's home city
- An Egyptian Chinese teacher gives each AI tool its own job
- Mistral Large 4 vs DeepSeek V4 Pro: Navigating the open model landscape
- Chinese AI firms accused of using OpenAI data to train new models
- Kimi and DeepSeek offer different trade-offs for Chinese AI users
- Chinese AI firms use OpenAI outputs to train rival models
- DeepSeek V5 names appear in test code as V4.1-Flash tops open models on coding
- Report says Chinese AI labs leaned on OpenAI's answers to train their models
- DeepSeek founder Liang Wenfeng enters Hurun's top 10 for the first time
- A sandboxed coding agent can switch off its own sandbox in DeepSeek Harness
- DeepSeek V5 code names show up in the company's official repository
- AI chatbots show diverging accuracy in male infertility answers
- Research highlights safety gaps in DeepSeek's latest AI model
- DeepSeek AI model shows safety gaps in independent tests
- Tests show DeepSeek model can bypass its own safety software
- DeepSeek founder Liang Wenfeng enters China's rich list with $41.7B fortune
- DeepSeek eyes $15 billion funding round as valuation hits $74 billion
- DeepSeek is hiring hard as a huge new funding round nears, sources say
- DeepSeek and Huawei collaborate to close the gap with Nvidia hardware
- DeepSeek Qwen and Mistral compete for developer budgets with new models
- Chinese AI firms linked to using pirated Korean books for training
- AWS says its cloud capacity is sold out through 2027 as Chinese AI labs register to export models
- DeepSeek and MiniMax finish registration for sending AI models abroad
- China lets DeepSeek and MiniMax keep selling AI abroad under new export rules
- DeepSeek reportedly raises $15 billion in massive funding round
- China grants AI export licenses to DeepSeek and MiniMax
- Open-weight DeepSeek model outperforms Google's gated Gemini in bug hunting
- vLLM says its open engine is within 6% of DeepSeek's in-house DeepInfer
- China reportedly clears DeepSeek and MiniMax to deploy AI models overseas
- Chinese AI firms DeepSeek and MiniMax gain global market access
- US ban on DeepSeek may deter private firms from using low-cost Chinese AI
- Listed Chinese firms back DeepSeek's record round, mostly through funds
- A Chinese commentator says DeepSeek has exposed a US AI bubble. Treat it as opinion
- China approves DeepSeek and MiniMax AI models for global use
- DeepSeek and MiniMax register under China's draft AI export controls
- China approves DeepSeek and MiniMax for overseas AI deployment
- DeepSeek raises $15bn from Tencent and CATL ahead of IPO
- DeepSeek secures 1 trillion RMB in record-breaking industrial funding
- DeepSeek and MiniMax clear China's new AI export registration
- DeepSeek raises €13 billion to challenge US AI leaders
- DeepSeek AI model creates exploit to bypass Chrome sandbox security
- DeepSeek and MiniMax receive approval for overseas AI deployment
- Fireworks says DeepSeek-V4.1-Flash matches Astra on coding tests for far less
- Chinese models reportedly carry 55% of Vercel's enterprise AI traffic
- Chinese open-source models capture over half of enterprise AI traffic
- Chinese models reportedly carry 55% of Vercel's enterprise AI traffic
- DeepSeek raises $15 billion in funding despite US security restrictions
- DeepSeek reportedly raises $15 billion in latest funding round
- DeepSeek raises nearly $15 billion to fuel Chinese AI expansion
- DeepSeek raises $15 billion to compete with US AI leaders
- DeepSeek raises nearly $15 billion to accelerate Chinese AI development
- DeepSeek raises 80 billion yuan to challenge US AI leaders
- Nvidia's Jensen Huang defends GPU lifespans, CoreWeave financing and AI demand
- Tencent's AI chief wants his team to work hard enough to make DeepSeek work overtime
- China demands the US revoke its federal ban on DeepSeek
- DeepSeek reportedly raises about 100 billion yuan, with CATL and Tencent leading