Kimi K2 ThinkingOpen Weight
Released
Kimi K2 Thinking is an AI model released by Moonshot AI on Thursday, Nov 6 2025, 7 days after Kimi Linear. It is an open-weight model — the trained weights are available to download and run. It is a 1T parameterThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. model with a 256k token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider.. Benchmark results (shown below) cover SWE-Bench Verified.
Get Moonshot AI releases and AI news.
Every other Monday
Available from
Benchmarks
Compare Kimi K2 Thinking with
Suggested comparisons
Kimi K2 ThinkingvsKimi K3.1 ProKimi K2 ThinkingvsGPT-6.1 TerraKimi K2 ThinkingvsClaude Haiku 5.5Kimi K2 ThinkingvsGemini 4 CarbonKimi K2 ThinkingvsMuse Spark OpenKimi K2 ThinkingvsGrok 4.7Kimi K2 ThinkingvsDeepSeek-V4.1-Flash-1017Kimi K2 ThinkingvsMistral Large 4Kimi K2 ThinkingvsGLM-5.3-FlashKimi K2 ThinkingvsQwen3.8-Max-0902Kimi K2 ThinkingvsNemotron 3.5 Bolt E4B 4GBAbout
Kimi K2 Thinking was introduced in November 2025 as a reasoning-oriented member of Moonshot’s K2 family. The mixture-of-experts model had one trillion total parameters with about 32 billion active per token and supported a 256K-token context.
The model was designed to interleave reasoning with tool calls across extended tasks, including coding and search workflows. Moonshot published downloadable weights and documented the inference settings used for its evaluations. Those agent results depended on tools and orchestration as well as the model, so they should be compared under the stated setup.
Frequently asked questions
Kimi K2 Thinking was released by Moonshot AI on Thursday, Nov 6 2025.