SWE-bench Verified (Vals run)
Human-verified GitHub issues from Python repositories, bucketed by fix time. Close to saturation; shares weight with other SWE-bench Verified runs. Run by Vals AI.
| Category | Coding |
|---|---|
| Metric | Accuracy (%) |
| Source | Vals AI (evaluator) |
| License | None stated |
| Benchmark page | www.vals.ai |
| Status | Scored |
| Weight | 0.8 within Coding |
| Fit | Discrimination 1.04, noise 0.40 |
| Lineage | Shares weight with swe_bench_verified |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Claude Opus 5 | Anthropic | 97.0% | 2026-09-01 |
| 2 | DeepSeek V4 Pro 0813 Max | DeepSeek | 96.4% | 2026-09-01 |
| 3 | GPT-5.6 Sol Max | OpenAI | 96.2% | 2026-09-01 |
| 4 | Grok 4.6 High | xAI | 95.6% | 2026-09-01 |
| 5 | GLM-5.3 Max | Zhipu AI | 95.4% | 2026-09-01 |
| 6 | GPT-5.6 Terra Max | OpenAI | 95.4% | 2026-09-01 |
| 7 | Claude Fable 5 Max | Anthropic | 95.0% | 2026-09-01 |
| 8 | Kimi K3 | Moonshot AI | 93.4% | 2026-09-01 |
| 9 | GPT-5.6 Luna Max | OpenAI | 93.0% | 2026-09-01 |
| 10 | GLM-5.3-Flash Max | Zhipu AI | 92.0% | 2026-09-01 |
| 11 | DeepSeek V4 Flash 0731 High | DeepSeek | 88.8% | 2026-09-01 |
| 12 | Claude Opus 4.8 Max | Anthropic | 88.6% | 2026-09-01 |
| 13 | Grok 4.5 High | xAI | 86.6% | 2026-09-01 |
| 14 | Muse Spark 1.2 Extra High | Meta | 86.6% | 2026-09-01 |
| 15 | Qwen3.8 27B Extra High | Alibaba | 86.0% | 2026-09-01 |
| 16 | Claude Opus 4.8 | Anthropic | 85.8% | 2026-09-01 |
| 17 | Qwen3.8 Max | Alibaba | 85.6% | 2026-09-01 |
| 18 | GLM-5.2 Max | Zhipu AI | 82.8% | 2026-09-01 |
| 19 | GPT-5.5 Extra High | OpenAI | 82.6% | 2026-09-01 |
| 20 | Inkling Small | Thinking Machines | 82.2% | 2026-09-01 |
| 21 | Claude Opus 4.7 Max | Anthropic | 82.0% | 2026-09-01 |
| 22 | Muse Spark 1.1 Extra High | Meta | 82.0% | 2026-09-01 |
| 23 | Gemini 3.7 Flash High | 80.8% | 2026-09-01 | |
| 24 | Gemini 3.8 Flash High | 80.0% | 2026-09-01 | |
| 25 | Composer 2.5 | Cursor | 79.6% | 2026-09-01 |
| 26 | Claude Sonnet 5 Max | Anthropic | 79.6% | 2026-09-01 |
| 27 | Gemini 3.6 Flash High | 79.6% | 2026-09-01 | |
| 28 | Gemini 3.1 Pro Preview High | 78.8% | 2026-09-01 | |
| 29 | Gemini 3.5 Flash High | 78.8% | 2026-09-01 | |
| 30 | Kimi K2.7 Code | Moonshot AI | 78.2% | 2026-09-01 |
| 31 | GPT-5.4 Extra High | OpenAI | 78.2% | 2026-09-01 |
| 32 | Claude Opus 4.6 Thinking | Anthropic | 78.2% | 2026-09-01 |
| 33 | GPT-5.3 Codex Extra High | OpenAI | 78.0% | 2026-09-01 |
| 34 | Inkling | Thinking Machines | 77.6% | 2026-09-01 |
| 35 | Claude Sonnet 4.6 Max | Anthropic | 77.4% | 2026-09-01 |
| 36 | DeepSeek V4 Pro Max | DeepSeek | 77.4% | 2026-09-01 |
| 37 | GLM-5.1 | Zhipu AI | 76.4% | 2026-09-01 |
| 38 | Claude Opus 4.5 Thinking | Anthropic | 76.4% | 2026-09-01 |
| 39 | Gemini 3 Pro Preview High | 76.4% | 2026-09-01 | |
| 40 | Kimi K2.6 | Moonshot AI | 76.2% | 2026-09-01 |
| 41 | GPT-5.2 Extra High | OpenAI | 75.8% | 2026-09-01 |
| 42 | MiniMax-M3 | MiniMax | 75.0% | 2026-09-01 |
| 43 | Gemini 3.5 Flash Lite High | 75.0% | 2026-09-01 | |
| 44 | Gemini 3 Flash Preview High | 75.0% | 2026-09-01 | |
| 45 | MiniMax-M2.1 | MiniMax | 74.8% | 2026-09-01 |
| 46 | Muse Spark | Meta | 74.4% | 2026-09-01 |
| 47 | MiniMax-M2.5 | MiniMax | 74.2% | 2026-09-01 |
| 48 | MiMo-V2.5-Pro | Xiaomi | 74.0% | 2026-09-01 |
| 49 | MiniMax-M2.7 | MiniMax | 73.8% | 2026-09-01 |
| 50 | Qwen3.6 Plus | Alibaba | 73.4% | 2026-09-01 |
| 51 | GPT-5.4 mini Extra High | OpenAI | 73.0% | 2026-09-01 |
| 52 | Qwen3.6 Max Preview | Alibaba | 72.8% | 2026-09-01 |
| 53 | GPT-5.2 Codex High | OpenAI | 72.4% | 2026-09-01 |
| 54 | Grok 4.20 (Reasoning) | xAI | 72.2% | 2026-09-01 |
| 55 | Grok 4.3 High | xAI | 71.4% | 2026-09-01 |
| 56 | GLM-5 Thinking | Zhipu AI | 71.4% | 2026-09-01 |
| 57 | Qwen3.5 Plus Thinking | Alibaba | 71.2% | 2026-09-01 |
| 58 | MiMo-V2.5 | Xiaomi | 71.0% | 2026-09-01 |
| 59 | Qwen3.6 27B | Alibaba | 70.0% | 2026-09-01 |
| 60 | Claude Sonnet 4.5 Thinking | Anthropic | 70.0% | 2026-09-01 |
| 61 | Kimi K2.5 Thinking | Moonshot AI | 70.0% | 2026-09-01 |
| 62 | GPT-5.4 nano High | OpenAI | 69.8% | 2026-09-01 |
| 63 | GPT-5.1 High | OpenAI | 69.8% | 2026-09-01 |
| 64 | GLM-4.7 | Zhipu AI | 69.4% | 2026-09-01 |
| 65 | Nemotron 3 Ultra 550B A55B | NVIDIA | 69.0% | 2026-09-01 |
| 66 | GPT-5 High | OpenAI | 69.0% | 2026-09-01 |
| 67 | Qwen3.7 Max | Alibaba | 68.8% | 2026-09-01 |
| 68 | DeepSeek V3.2 High | DeepSeek | 67.6% | 2026-09-01 |
| 69 | Claude Haiku 4.5 Thinking | Anthropic | 66.6% | 2026-09-01 |
| 70 | Mistral Medium 3.5 High | Mistral | 66.4% | 2026-09-01 |
| 71 | Qwen3.5 Flash | Alibaba | 64.4% | 2026-09-01 |
| 72 | Devstral 2 | Mistral | 62.8% | 2026-09-01 |
| 73 | Gemini 3.1 Flash Lite Preview High | 62.8% | 2026-09-01 | |
| 74 | GPT-5 Mini High | OpenAI | 60.8% | 2026-09-01 |
| 75 | Kimi K2 Thinking | Moonshot AI | 60.2% | 2026-09-01 |
| 76 | Grok 4 | xAI | 57.8% | 2026-09-01 |
| 77 | Laguna M.1 | Poolside | 57.6% | 2026-09-01 |
| 78 | Laguna XS.2 | Poolside | 55.2% | 2026-09-01 |
| 79 | Gemini 2.5 Pro | 54.4% | 2026-09-01 | |
| 80 | X-Ai/Grok-4-Fast-Reasoning | xAI | 45.4% | 2026-09-01 |
| 81 | Grok 4.1 Fast (Reasoning) | xAI | 41.4% | 2026-09-01 |
| 82 | Mistral Large 3 | Mistral | 41.4% | 2026-09-01 |
| 83 | GPT OSS 120B | OpenAI | 33.6% | 2026-09-01 |
| 84 | Command A | Cohere | 7.8% | 2026-09-01 |