# MoonshotAI Kimi K2 Thinking

MoonshotAI Kimi K2 Thinking costs $0.60 per million input tokens and $2.50 per million output tokens, as of October 22, 2026. Only one of the 6 coding benchmarks behind the index has measured it, and the index needs two, so there is not enough data for a coding score. 29 benchmarks outside the index have a result below.

Page: https://llmbenchmark.ai/models/kimi-k2-thinking
Data from October 22, 2026. Every figure below names the board it was read from and the day.

- Maker: Moonshot AI
- Released: November 6, 2025
- Price per million input tokens: $0.60
- Price per million output tokens: $2.50
- Context window: 262K tokens
- Vibe Coding Index: not enough data (one index benchmark; at least two are needed to rank)

Benchmark results (30 of 195):

- SWE-bench Verified: 63.4% (source: SWE-bench (swebench.com), read October 22, 2026)
- GPQA Diamond: 78.5% (source: GPQA (Vals AI), read October 22, 2026)
- AIME: 85.4% (source: AIME (Vals AI), read October 22, 2026)
- MMLU Pro: 81.1% (source: MMLU Pro (Vals AI), read October 22, 2026)
- AA Intelligence Index: 22.0 (source: Artificial Analysis (via OpenRouter), read October 22, 2026)
- Design Arena: websites: 1113 Elo (source: Design Arena, read October 22, 2026)
- LiveCodeBench (Vals AI): 63.1% (source: Vals AI boards, read October 22, 2026)
- SWE-bench Verified (Vals AI): 60.2% (source: Vals AI boards, read October 22, 2026)
- MGSM (Vals AI): 90.2% (source: Vals AI boards, read October 22, 2026)
- LegalBench (Vals AI): 80.2% (source: Vals AI boards, read October 22, 2026)
- CaseLaw (Vals AI): 65.7% (source: Vals AI boards, read October 22, 2026)
- CorpFin (Vals AI): 60.6% (source: Vals AI boards, read October 22, 2026)
- TaxEval (Vals AI): 71.7% (source: Vals AI boards, read October 22, 2026)
- MedQA (Vals AI): 92.6% (source: Vals AI boards, read October 22, 2026)
- AA Math Index: 94.7 (source: Artificial Analysis, read October 22, 2026)
- GPQA Diamond (Artificial Analysis run): 83.8% (source: Artificial Analysis, read October 22, 2026)
- Humanity's Last Exam (Artificial Analysis run): 23.8% (source: Artificial Analysis, read October 22, 2026)
- MMLU Pro (Artificial Analysis run): 84.8% (source: Artificial Analysis, read October 22, 2026)
- LiveCodeBench (Artificial Analysis run): 85.3% (source: Artificial Analysis, read October 22, 2026)
- Terminal-Bench Hard (Artificial Analysis run): 31.1% (source: Artificial Analysis, read October 22, 2026)
- IFBench: 68.1% (source: Artificial Analysis, read October 22, 2026)
- Long context reasoning (AA-LCR): 72.0% (source: Artificial Analysis, read October 22, 2026)
- τ²-Bench Telecom: 93.0% (source: Artificial Analysis, read October 22, 2026)
- AIME 2025 (Artificial Analysis run): 94.7% (source: Artificial Analysis, read October 22, 2026)
- Terminal-Bench: 35.7% (source: Epoch AI benchmarking hub, read October 22, 2026)
- WeirdML: 42.8% (source: Epoch AI benchmarking hub, read October 22, 2026)
- METR time horizons: 59.2% (source: Epoch AI benchmarking hub, read October 22, 2026)
- CL-bench: 17.6% (source: Epoch AI benchmarking hub, read October 22, 2026)
- ALE-Bench: 597.5 (source: ALE-Bench (Sakana AI), read October 22, 2026)
- Epoch Capabilities Index: 145.8 (source: Epoch AI benchmarking hub, read October 22, 2026)