Benchmark
SWE-bench Multilingual
Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.
- Models measured
- 65
- Scores recorded
- 103
- Sources
- 3
- IQ weight
- 2.97
Score distribution
- 71.9%
- Mean
- 73.8%
- Median
- 13.1%
- Std dev
- 54.5%
- 10th percentile
- 86.6%
- 90th percentile
- 30.5%
- Min
This benchmark has a weight of 2.97 for chat models, measured over 58 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 96.6% |
| 2 | Claude Opus 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 93.9% |
| 3 | Claude Opus 5.5Opus 5.5 System Card capability section· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 93.9% |
| 4 | Claude Sonnet 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 90.3% |
| 5 | Claude Sonnet 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 90.3% |
| 6 | Claude Opus 5Opus 5 System Card capability summary · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 89.5% |
| 7 | Claude Opus 5Opus 5.5 System Card capability section· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 89.5% |
| 8 | Claude Opus 5System Card capability summary · adaptive thinking · max effort · default sampling · 5 trials · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 89.5% |
| 9 | Claude Fable 5.1Opus 5.5 System Card capability section· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 89.1% |
| 10 | Claude Fable 5.1System Card capability summary · adaptive thinking · max effort · default sampling · 5 trials · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 89.1% |
| 11 | Ling 3.1 Flash· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 88.5% |
| 12 | Claude Fable 5Opus 5 System Card capability summary · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 86.6% |
| 13 | Claude Fable 5System Card capability summary · adaptive thinking · max effort · default sampling · 5 trials · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 86.6% |
| 14 | Claude Opus 4.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.4% |
| 15 | Claude Opus 4.8Opus 4.8 System Card capability summary · adaptive thinking · max effort · 5 trials unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.4% |
| 16 | Claude Opus 4.8Opus 5 System Card capability summary · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.4% |
| 17 | Claude Haiku 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.7% |
| 18 | Claude Haiku 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 83.7% |
| 19 | Hy4 preview· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 82.9% |
| 20 | Hy4 previewHF Eval Result / Resolved· Reported by Tencent Hy· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 82.9% |
| 21 | Inkling 2 Small· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.2% |
| 22 | Qwen 3.8 Flash Next· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 81% |
| 23 | Qwen 3.8 Flash NextQwen3.8-Flash-Next release/model card · mini-SWE-agent · temp 1.0 · top_p 0.95 · 256K· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81% |
| 24 | Qwen3.8 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81% |
| 25 | Qwen3.8 Flashmini-SWE-agent · temp 1.0 · top_p 0.95 · 256K· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 81% |
| 26 | Claude Opus 4.7adaptive thinking · max effort · avg@5· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 80.5% |
| 27 | Claude Opus 4.7Opus 4.8 System Card capability summary · adaptive thinking · max effort · 5 trials unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 80.5% |
| 28 | Qwen3.8 Omni Flashmini-SWE-agent · temp 1.0 · top_p 0.95 · 256K· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 80.5% |
| 29 | Laguna S 2.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 78.5% |
| 30 | Beam 1.0 501B A23B· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 78.3% |
| 31 | Claude Sonnet 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 78.3% |
| 32 | Claude Sonnet 5adaptive thinking · max effort · 5 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 78.3% |
| 33 | Qwen3.7 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 78.3% |
| 34 | Qwen3.7 MaxQwen3.7-Max release · internal bash + file-edit agent scaffold · temp 1.0 · top_p 0.95 · 200K· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 78.3% |
| 35 | Claude Opus 4.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 77.8% |
| 36 | Claude Opus 4.6Anthropic Opus 4.6 launch page · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 77.8% |
| 37 | Claude Opus 4.6mini-SWE-agent · temp 1.0 · top_p 0.95 · 256K · max effort· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 77.5% |
| 38 | LongCat 2.0· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 77.3% |
| 39 | Kimi K2.6· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 76.7% |
| 40 | Kimi K2.6Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · in-house SWE-agent-derived harness · avg@10· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 76.7% |
| 41 | MiniMax M2.7· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 76.5% |
| 42 | MiniMax M2.7MiniMax M2.7 official release prose· Reported by MiniMax· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 76.5% |
| 43 | Claude Opus 4.564K thinking budget · 200K context · avg@5· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 76.2% |
| 44 | Claude Opus 4.5no extended thinking · 200K context · avg@5· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 76.2% |
| 45 | DeepSeek V4 Pro· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 76.2% |
| 46 | Claude Sonnet 4.6adaptive thinking · max effort · 10 trials· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 75.9% |
| 47 | Hy3 (Deep Infra)· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 75.8% |
| 48 | Hy3 (Deep Infra)Hy3 official Benchmark Appendix · highest reasoning tier unless noted· Reported by Tencent Hunyuan· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 75.8% |
| 49 | Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 75.8% |
| 50 | Qwen3.7 Plusmini-SWE-agent · temp 1.0 · top_p 0.95 · 256K· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 75.8% |
Showing 1–50 of 103 results.
Data credits: Hugging Face official benchmark leaderboards; licence varies by benchmark · Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.