DeepSeek AI News
Latest news, updates, and announcements from DeepSeek AI.
DeepSeek Launches V4.1-Flash at $0.003 per Million Off-Peak Cached Tokens, Claims Benchmark Edge Over GPT-5.6 and Claude Opus 5
DeepSeek has released V4.1-Flash, a new model priced at $0.003 per million tokens for off-peak cached input. The company says its benchmark results surpass OpenAI's GPT-5.6 and Anthropic's Claude Opus...
DeepSeek Hires 150 Senior Backend Engineers as It Shifts From Research Lab to Service Platform
DeepSeek's AI agent harness team lead, Cui Tianyi — a former quantitative trader who joined the company in March — said via social media on September 8 that DeepSeek is recruiting around 150 senior ba...
DeepSeek V4 Pro Deep-Dive Finds Post-Training Gains, Extrapolated 1M Context, Kimi K3 Parity
An independent review by Chinese outlet Leiphone tested DeepSeek's V4 Pro model and its open-source Harness agent framework — benchmark probes, seven coding-task runs, a 920,000-token long-context tes...
DeepSeek Seeks $7.4B Round at $74B Valuation, Opens Direct Equity Stakes Ahead of 2027 IPO
DeepSeek is reportedly pursuing a new funding round of $7.4 billion, per the Wall Street Journal, that would lift its valuation to $74 billion from more than $50 billion in June. Proceeds are earmarke...
DeepSeek's V4 Flash model, now ranking first on AI leaderboards, completed only 53.8% of real-world...
DeepSeek's V4 Flash model, now ranking first on AI leaderboards, completed only 53.8% of real-world agent tasks in a new evaluation. Concurrently, the company has raised API prices by up to 1100%, sig...
DeepSeek, the Chinese AI lab known for its low-cost models, has announced significant API price incr...
DeepSeek, the Chinese AI lab known for its low-cost models, has announced significant API price increases following the launch of its V4 Flash model on July 31, 2026. The company added a price adjustm...
Alibaba has released Qwen 3.8 27B, an Apache 2.0-licensed open-weight dense model with 27 billion pa...
Alibaba has released Qwen 3.8 27B, an Apache 2.0-licensed open-weight dense model with 27 billion parameters, designed for local execution on high-end GPUs and AMD Ryzen AI Max PCs. It natively suppor...
DeepSeek has announced significant increases to its API pricing, with some rates rising by up to 12...
DeepSeek has announced significant increases to its API pricing, with some rates rising by up to 12 times, and will implement peak-time pricing starting August 17. The new rates for DeepSeek-V4-Flash...
DeepSeek officially launched V4-Pro-0813, its flagship model now available via app, web, and API. Th...
DeepSeek officially launched V4-Pro-0813, its flagship model now available via app, web, and API. The company said the new version is significantly improved at supporting AI agents, a growing area of...
DeepSeek has reportedly selected 17 open-source projects for its Harness incubation program, with a...
DeepSeek has reportedly selected 17 open-source projects for its Harness incubation program, with a leaked partial list showing a clear focus on agent execution, routing, and orchestration rather than...
DeepSeek has introduced Harness, an open-source tool designed to compete with Anthropic's Claude Cod...
DeepSeek has introduced Harness, an open-source tool designed to compete with Anthropic's Claude Code for AI agent development, along with V4-Pro on its API featuring tiered pricing. This expansion si...
DeepSeek V4 Pro (DeepSeek-V4-Pro-0813) has gone live on the company's API, bringing enhanced agent c...
DeepSeek V4 Pro (DeepSeek-V4-Pro-0813) has gone live on the company's API, bringing enhanced agent capabilities that the company says approach or exceed those of Anthropic's Fable 5. The model support...
DeepSeek has released DeepSeek-V4-Pro-0813, the latest iteration of its flagship foundation model, d...
DeepSeek has released DeepSeek-V4-Pro-0813, the latest iteration of its flagship foundation model, drawing mixed results across key industry benchmarks. The release follows recent API price adjustment...
DeepSeek Announces Significant API Price Increases
DeepSeek announced that it will significantly raise prices for its API services. The notice warns that the increase could be substantial, but does not specify the new rates or an effective date. Curre...
DeepSeek announced a "significant" price increase for its API services, citing surging global demand...
DeepSeek announced a "significant" price increase for its API services, citing surging global demand for its ultra-cheap models. The Hangzhou-based startup released DeepSeek-V4-Flash-0731, a 284-billi...
DeepSeek announced on Thursday that it plans to raise prices across its API services, with the compa...
DeepSeek announced on Thursday that it plans to raise prices across its API services, with the company saying a relatively large increase is expected and that it will announce the final pricing scheme...
DeepSeek reportedly reopens talks on a RMB50 billion second funding round at a ~RMB500 billion valuation
DeepSeek has reportedly reopened discussions for a second funding round targeting RMB50 billion, which could value the Chinese AI company at approximately RMB500 billion before financing. The talks ma...
DeepSeek has released the formal version of its V4-Flash API into public beta, rolling out an upgrad...
DeepSeek has released the formal version of its V4-Flash API into public beta, rolling out an upgrade centered on agent task performance. The company reports that the V4-Flash-0731 model achieved scor...
DeepSeek, the Chinese AI startup known for its low-cost foundation model, previewed a new model spec...
DeepSeek, the Chinese AI startup known for its low-cost foundation model, previewed a new model specifically optimized to run on Huawei's AI chips. The adaptation marks a deepening integration between...
DeepSeek has paused its second fundraising round, a deal that would have valued the Chinese AI lab a...
DeepSeek has paused its second fundraising round, a deal that would have valued the Chinese AI lab at approximately $71 billion, after founder Liang Wenfeng's frustrations over leaked comments from a...