Kimi K2 Thinking
kimi-k2-thinkingKimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning. Built on the trillion-parameter Mixture-of-Experts (MoE) architecture introduced in... Built by MoonshotAI.
Prices updated
Input price
$0.60
per 1M tokens · Standard
Output price
$2.50
per 1M tokens · Standard
Input limit
262K
tokens
Output limit
236K
tokens
Input formats
Output formats
Kimi K2 Thinking price history
2 price records since 9/25/2026
Kimi K2 Thinking cost calculator
$25/month
Standard pricing
Overview
What is Kimi K2 Thinking?
Kimi K2 Thinking is a text & reasoning model from MoonshotAI. Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning. Built on the trillion-parameter Mixture-of-Experts (MoE) architecture introduced in... Its 262K context window and $0.60 input price make it a candidate for quality-focused production applications.
Benchmarks
Kimi K2 Thinking benchmarks & speed
How Kimi K2 Thinking scores on standardized evaluations, and where it lands among every model we track.
22.0
Intelligence Index
156tok/s
Output speed
Reasoning
GPQA Diamond
83.8%
Graduate-level scientific reasoning · top 48%
HLE
23.8%
Humanity's Last Exam · top 46%
Latency & design
- Time to first token
- 1.3s
Independent scores from Artificial Analysis and DesignArena · updated 10/5/2026. Higher is better; ranks compare against every model we track with that score.
Rates
Kimi K2 Thinking pricing
Pricing is based on token usage. Provider-specific caching, batch, regional, and tool charges may affect the final cost.
Input · Standard
$0.60
Per 1M tokens
Cached input · Standard
$0.15
Per 1M tokens
Capabilities
Kimi K2 Thinking Tools
Tools available when using Kimi K2 Thinking through supported provider APIs.
Function calling
Supported
Structured outputs
Supported
JSON mode
Supported
Reasoning
Supported
Built-in web search
Not supported
Log probabilities
Supported
Deterministic seed
Supported
Parallel tool calls
Not supported
Prompt caching
Not supported
Strengths and limitations
Strengths
- Measured results (AA Intelligence 22) position it for demanding production work.
- 262K context supports large documents, repositories and extended conversations.
- Supports text & reasoning workloads in one model.
Limitations
- Generated tokens cost 4× more than input tokens, which matters for verbose responses.
- Large context capacity does not guarantee consistent retrieval across the entire prompt.
- Arena Elo and MMLU-Pro are directional; test accuracy, latency and reliability on your own workload before committing.
Arena Elo, MMLU-Pro and pricing figures are illustrative; validate current vendor terms before purchase.
Same provider