M
Kimi K2 ThinkingOPEN
Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning.
Prices update automatically — checked daily against provider list prices.
- $1.08 per 1M tokens, assuming 3 input for 1 output the convention used by Artificial Analysis
- cheaper than 14% of the 191 open-weight models tracked here
- 3 of 7 Moonshot AI models tracked here, from cheapest
Benchmarks & performance
Independent benchmark scores for Kimi K2 Thinking, measured by Artificial Analysis. Higher is better.
For the same budget, Kimi K2 Thinking delivers 5 times more capability. Open weights also mean you can self-host it and pay nothing per token.
Models at this level cost $0.74 per 1M tokens (median of 22) — this one costs $1.08.
Cheaper and better on this index: GLM 5.3 Flash · Gemini 3.7 Flash · Gemini 3.7 Flash (batch) and 34 more
About this model
Kimi K2 Thinking is an open-weight AI model by Moonshot AI. You can download and self-host it for free; the prices below are hosted-API list prices, tracked daily, for when you prefer convenience over self-hosting.
Frequently asked questions
What is Kimi K2 Thinking?
Kimi K2 Thinking is an AI language model from Moonshot AI. It is open-weight: you can download it and run it on your own hardware, for free. It scores 33.5 on the Artificial Analysis intelligence index.
Is Kimi K2 Thinking free?
The weights are free and open — you can self-host Kimi K2 Thinking and pay nothing per token. If you prefer a hosted API, list prices are $0.6 per million input tokens and $2.5 per million output tokens.
What is Kimi K2 Thinking good at?
Independent benchmarks from Artificial Analysis give it GPQA 83.8%, MMLU-Pro 84.8%, Humanity's Last Exam 23.8%, Long Context Reasoning 70.3%, LiveCodeBench 85.3%, SciCode 42.4%, AIME 2025 94.7%, IFBench 68.1%, τ²-Bench 93%, Terminal-Bench Hard 31.1%. It is particularly used for mathematical reasoning.
Can I self-host Kimi K2 Thinking?
Yes. Kimi K2 Thinking has open weights, so you can download it and run it on your own GPU or server with tools like Ollama, vLLM or llama.cpp — with no per-token cost.