New models, open releases and updates: flagship launches, open weights, and price and performance changes as they happen.
DeepSeek V2 Chat and DeepSeek Coder V2 have been merged and upgraded into DeepSeek V2.5. API users reach the new model through either deepseek-coder or deepseek-chat.
Why it matters: DeepSeek merged its chat and code models into V2.5 while keeping both existing API names so current call patterns still work.
The deepseek-chat model has been upgraded to DeepSeek-V2-0628. Reasoning improved, and role-play got a clear boost. HumanEval Pass@1 rose from 79.88% to 84.76%, MATH ACC@1 from 55.02% to 71.02%, and BBH from 78.56% to 83.40%.
Why it matters: The math benchmark gain is larger than the code and reasoning gains, and the win rate against GPT-4-0314 on Arena-Hard also improved.
The deepseek-coder model has been upgraded to DeepSeek-Coder-V2-0614, with a clear gain in coding ability. The company says its code generation, code understanding, code repair and code completion reach the level of GPT-4-Turbo-0409, with strong math and reasoning, while general ability matches DeepSeek-V2-0517.
Why it matters: DeepSeek upgrades its code model and benchmarks code generation, understanding, repair and completion against GPT-4-Turbo-0409.