Kimi K2.7 Code model card (Hugging Face) Jun 12, 2026
- Kimi Code Bench v2 62
- Program Bench
- MLS Bench Lite 35.1
- Kimi Claw 24/7 Bench 46.9
- MCP Atlas 76
- MCP Mark Verified 81.1
Kimi K2.7 Code
Kimi K2.7 Code ranks first on LiveClawBench.
74
Benchmarks with a score
1
Ranked first, out of five or more models
2
In the top three
Moonshot AI Jun 2026
Kimi K2.7 Code
Open weights Mid-size
151.2
#64 of 256
Epoch Capabilities Index
- DeepSWE 30.5% #26 /28
- AutomationBench 3.20% #37 /42
- GPQA Diamond 87.9% #66 /229
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DesignArena ASCII Art | 1,185 | #35 of 77 | Official leaderboard | |
| SimpleQA Verified | 36.5% | #59 of 94 | Epoch AI run | |
| AA-LCR Extended | 49.2% | #35 of 49 | Official leaderboard | |
| HLE Extended | 13.7% | #23 of 31 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 95.6% | #49 of 212 | Epoch AI run | |
| GPQA Diamond | 87.9% | #66 of 229 | Epoch AI run | |
| SimpleBench | 57.9% | #48 of 106 | Official leaderboard | |
| FrontierMath Tiers 1-3 v2 | 54% | #50 of 86 | Official leaderboard | |
| Surface Evolver Bench | 48.8% | #16 of 25 | Official leaderboard | |
| FrontierMath Tier 4 v2 | 12.2% | #56 of 68 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DesignArena Website | 1,259 | #47 of 151 | Official leaderboard | |
| DesignArena UI Component | 1,252 | #50 of 139 | Official leaderboard | |
| WeirdML | 54.1% | #41 of 112 | Official leaderboard | |
| DesignArena 3D Design | 1,223 | #50 of 136 | Official leaderboard | |
| DesignArena Code | 1,244 | #54 of 143 | Official leaderboard | |
| DesignArena Game Dev | 1,210 | #59 of 143 | Official leaderboard | |
| DesignArena Data Visualization | 1,222 | #55 of 118 | Official leaderboard | |
| WebDev Arena | 1,473 | #50 of 105 | arena.ai, via Epoch AI | |
| AndroidBench | 70.4% | #16 of 32 | Official leaderboard | |
| Code Arena | 1,474 | #66 of 128 | Official leaderboard | |
| ALE-Bench | 1,422 | #65 of 126 | Official leaderboard | |
| DesignArena SVG | 1,166 | #54 of 103 | Official leaderboard | |
| APEX-SWE | 40.3% | #35 of 65 | Official leaderboard | |
| DesignArena Full Stack | 1,154 | #27 of 49 | Official leaderboard | |
| DAXBench v2 | 77.8% | #26 of 47 | Official leaderboard | |
| Vibe Code Bench | 47.2% | #70 of 120 | Official leaderboard | |
| DesignArena Agentic Overall | 1,151 | #44 of 74 | Official leaderboard | |
| Code Migration | 25.4% | #48 of 78 | Official leaderboard | |
| Image-to-WebDev Arena | 1,512 | #36 of 55 | Official leaderboard | |
| DesignArena Image to Frontend | 1,149 | #21 of 31 | Official leaderboard | |
| DesignArena Frontend Web App | 1,158 | #36 of 53 | Official leaderboard | |
| DesignArena Video to Website | 1,175 | #13 of 19 | Official leaderboard | |
| DesignArena Image to HTML | 1,181 | #35 of 51 | Official leaderboard | |
| Claw-SWE-Bench | 68.3% | #11 of 16 | Official leaderboard | |
| DesignArena React Native | 1,136 | #39 of 56 | Official leaderboard | |
| MLS-Bench | 35.1% | #11 of 15 | Official leaderboard | |
| SWE Atlas Codebase QnA Extended | 4.70% | #28 of 38 | Official leaderboard | |
| DesignArena Android (Kotlin) | 1,129 | #34 of 46 | Official leaderboard | |
| DesignArena Godot Game Dev | 1,128 | #18 of 24 | Official leaderboard | |
| EvoCode-Bench v2 | 7.80% | #10 of 13 | Official leaderboard | |
| ReactBench | 23.6% | #23 of 28 | Official leaderboard | |
| SWE-bench Verified Extended | 26.7% | #37 of 44 | Official leaderboard | |
| DesignArena Agentic Game Dev | 1,078 | #23 of 27 | Official leaderboard | |
| GBAEval | 0.85% | #19 of 21 | Official leaderboard | |
| DeepSWE | 30.5% | #26 of 28 | Official leaderboard | |
| Next.js Agent Evals | 75% | #14 of 14 | Official leaderboard | |
| Kimi Code Bench v2 | 62% | #3 of 3 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveClawBench | 76% | #1 of 17 | Official leaderboard | |
| MCPMark | 81.1% | #2 of 10 | Lab model card | |
| Pencil Puzzle Bench | 16.7% | #16 of 53 | Official leaderboard | |
| PinchBench 2.0 | 76.1% | #20 of 57 | Official leaderboard | |
| MCP Atlas | 76% | #21 of 46 | Lab model card | |
| Long-Horizon-Terminal-Bench (LHTB) | 36.7% | #14 of 25 | Official leaderboard | |
| Toolathlon-Verified | 58% | #14 of 25 | Official leaderboard | |
| WildClawBench | 46.9% | #18 of 31 | Official leaderboard | |
| MCPMark Verified | 81.9% | #5 of 8 | Official leaderboard | |
| RuneBench | 4.18 | #52 of 76 | Official leaderboard | |
| Terminal-Bench 2.1 Extended | 23.6% | #34 of 44 | Official leaderboard | |
| Chess Puzzles | 14% | #81 of 93 | Epoch AI run | |
| AutomationBench | 3.20% | #37 of 42 | Official leaderboard | |
| BrowseComp Extended | 8.00% | #37 of 42 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DesignArena HTML Slides | 1,096 | #13 of 28 | Official leaderboard | |
| Vending-Bench 2 | $5,083 | #36 of 67 | Official leaderboard | |
| MedXpertQA MM Extended | 51.6% | #27 of 41 | Official leaderboard | |
| DesignArena Python-PPTX Slides | 1,112 | #26 of 36 | Official leaderboard | |
| APEX-1 | 60.1% | #52 of 71 | Official leaderboard | |
| APEX-Agents 1.1 | 53.4% | #54 of 65 | Official leaderboard | |
| APEX-Agents | 37.6% | #51 of 61 | Collected by Epoch AI, source not named | |
| GDPval Extended | 0.20% | #37 of 43 | Official leaderboard | |
| DAYJOB: Finance | 0% | #37 of 39 | Official leaderboard | |
| DAYJOB: Healthcare | 0% | #36 of 37 | Official leaderboard | |
| Kimi Claw 24/7 Bench | 46.9% | #1 of 1 | Lab model card |
Vision, audio and robotics
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMMU-Pro Extended | 23.1% | #34 of 38 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DrugDiscoveryBench | 35.3% | #14 of 20 | Official leaderboard |
Its lab's numbers next to 2 leaderboards
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
Moonshot AI's models over time
Moonshot AI's models in color, against every other model on the Epoch Capabilities Index, by release date.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.