Moonshot AI
Kimi K2 Thinking benchmark results
11 results on 7 benchmarks from 4 sources.
Also listed as: kimi-k2-thinking-turbo · Kimi K2 Thinking Turbo · kimi-k2-thinking · kimi-k2-0905-preview · Kimi K2 · Lingxi v1.5 x Kimi K2 · Kimi-K2-Thinking · Kimi-k2
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1430 | 114 | Overall, style control on; 63390 votes; 95% CI 1427–1433“kimi-k2-thinking-turbo” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1322 | 119 | Overall; 13093 votes; 95% CI 1315–1329“kimi-k2-thinking-turbo” | Official leaderboard | Oct 8, 2026 |
| GPQA Diamond | 84.2% | 91 | Epoch AI run, accuracy | Independent run | Nov 11, 2025 |
| SWE-bench Verified | 71.2% | 22 | agent: Lingxi v1.5“kimi-k2-0905-preview” | Official leaderboard | Oct 14, 2025 |
| SWE-bench Verified | 63.4% | 43 | agent: mini-SWE-agent“Kimi-K2-Thinking” | Official leaderboard | Dec 10, 2025 |
| AIME (OTIS Mock AIME 2024–2025) | 83.1% | 104 | Epoch AI run, OTIS Mock AIME 2024–2025 | Independent run | Nov 13, 2025 |
| τ²-bench | 56.5% | 18 | airline domain, pass^1; submitted by Moonshot AI; v0.1.3“Kimi-k2” | Vendor-reported | Jul 28, 2025 |
| τ²-bench | 70.6% | 24 | retail domain, pass^1; submitted by Moonshot AI; v0.1.3“Kimi-k2” | Vendor-reported | Jul 28, 2025 |
| τ²-bench | 65.8% | 16 | telecom domain, pass^1; submitted by Moonshot AI; v0.1.3“Kimi-k2” | Vendor-reported | Jul 28, 2025 |
| Terminal-Bench 2.0 | 35.7% | 110 | Terminal-Bench 2.0; agent: Terminus 2 | Official leaderboard via Epoch AI | Nov 6, 2025 |
| Terminal-Bench 2.0 | 35.7% | 111 | Terminal-Bench 2.0; agent: Terminus 2 | Official leaderboard via Epoch AI | Nov 11, 2025 |
FAQ
Where do these Kimi K2 Thinking scores come from?
From Epoch AI Benchmarking Hub, LMArena leaderboard dataset, SWE-bench leaderboards (swebench.com), and τ²-bench leaderboard submissions. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.