Latest / Frontier models and safety

DeepSeek releases R1 reasoning model with open weights under MIT licence

Model releaseFrontierComputeCNConfirmed

Chinese lab DeepSeek released DeepSeek-R1, a reasoning model trained largely with reinforcement learning that it said matches OpenAI o1 on maths, code and reasoning. Weights were published under the MIT licence alongside six smaller distilled models, and API output was priced at $2.19 per million tokens.

Why it matters

It showed that near-frontier reasoning could be reproduced cheaply and given away openly, resetting price expectations and sharpening US debate on export controls and open weights.

SourceDeepSeek API Docs Checked against the primary source. Independently fact-checked on 7 Oct 2026.
DeepSeekOpenAI

Line of Thought

Follow this story

Pick any item to keep going. Your path builds up above as a line you can share.

Directly linked

Connections our researchers recorded

What happened next

Later developments on the same thread

Same story elsewhere

What other countries and bodies did on this

Threads by topic: Model releases Open-weight models Frontier models Export controls