Retour à Deepseek V3
DeepSeek's updated 671B MoE chat model — improved instruction-following and reasoning over V3, still at open-weight cost.
164K tokensMoEProprietary
Benchmarks
DeepSeek-V3.2
70.0%DeepSeek V3.2 high reasoning — mini-SWE-agent v2
DeepSeek's updated 671B MoE chat model — improved instruction-following and reasoning over V3, still at open-weight cost.
DeepSeek V3.2 high reasoning — mini-SWE-agent v2