Benchmark
SWE-bench Verified
Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.
- Models measured
- 130
- Scores recorded
- 205
- Sources
- 5
- IQ weight
- 2.70
Score distribution
- 70%
- Mean
- 73.5%
- Median
- 13.7%
- Std dev
- 51.6%
- 10th percentile
- 80.8%
- 90th percentile
- 8.7%
- Min
This benchmark has a weight of 2.70 for chat models, measured over 129 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 |
|---|---|---|
| 101 | Seed 2.0 Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.5% |
| 102 | MiMo V2 Flash· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 73.4% |
| 103 | MiMo V2 FlashMiMo-V2-Flash final evaluation · published code-agent setting · harness unspecified· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 73.4% |
| 104 | Qwen3.6 35B A3B· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 73.4% |
| 105 | Qwen3.6 FlashQwen3.6-35B-A3B official model card · hosted API name Qwen3.6-Flash· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 73.4% |
| 106 | Claude Haiku 4.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.3% |
| 107 | Claude Haiku 4.5launch-page setting· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 73.3% |
| 108 | DeepSeek Reasoner· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.1% |
| 109 | DeepSeek V3.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 73.1% |
| 110 | DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · DeepSeek internal code-agent framework· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 73.1% |
| 111 | DeepSeek V3.2 Speciale· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 73.1% |
| 112 | Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.9% |
| 113 | Claude Sonnet 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.7% |
| 114 | Claude Sonnet 4pass@1 · bash/editor tools · no extended thinking · avg@10 · top_p 0.95· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.7% |
| 115 | MAI-Code-1.1-Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.6% |
| 116 | Claude Opus 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.5% |
| 117 | Claude Opus 4pass@1 · bash/editor tools · no extended thinking · avg@10 · top_p 0.95· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.5% |
| 118 | Qwen3.5 27B· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 3 sources
| 72.4% |
| 119 | Devstral 2 123b Instruct 2512official model card· Reported by Mistral AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.2% |
| 120 | ling-2.6-1t· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 72.2% |
| 121 | DeepSeek V3.2Non-thinking mode · temperature 1.0 · 128K context · resolved· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 72.1% |
| 122 | Qwen3.5 122B-A10B· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 3 sources
| 72% |
| 123 | Nemotron 3 Ultra· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 3 sources
| 71.9% |
| 124 | Kimi K2 Thinking· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 71.3% |
| 125 | NVIDIA Nemotron 3.5 Super VL 120B· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 71.1% |
| 126 | Laguna XS 2.1· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 70.9% |
| 127 | Grok Code Fast 1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.8% |
| 128 | Nemotron 3 UltraNeMo Evaluator SDK · official table· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 70.7% |
| 129 | Qwen3 Coder Next· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 70.6% |
| 130 | Solar Pro 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.6% |
| 131 | Claude Sonnet 3.7· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70.3% |
| 132 | LongCat-Flash-Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 70% |
| 133 | Laguna XS.2· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 69.9% |
| 134 | Qwen3 Coder· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 69.6% |
| 135 | Qwen3 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 69.6% |
| 136 | MiniMax M2· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 69.4% |
| 137 | Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 69.2% |
| 138 | OpenAI o3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 69.1% |
| 139 | o4 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 68.1% |
| 140 | GLM 4.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 68% |
| 141 | DeepSeek V3.2 Exp· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.8% |
| 142 | North Mini Code 1.0· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 67.6% |
| 143 | Gemini 2.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 67.2% |
| 144 | DeepSeek V3.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 66% |
| 145 | Kimi K2 Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 65.8% |
| 146 | Nova 2 Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 64.5% |
| 147 | Nova 2 LiteAmazon Nova 2 Lite official launch chart / technical report · inference-time scaling · internal agentic scaffold, v1· Reported by Amazon Web Services, Amazon AGI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 64.5% |
| 148 | GLM 4.5· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claimValues from 2 sources
| 64.2% |
| 149 | Trinity Large ThinkingTrinity-Large-Thinking Model Card · mini-swe-agent-v2· Reported by Arcee AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 63.2% |
| 150 | openPangu 2.0 Flash· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim | 63.1% |
Showing 101–150 of 205 results.
Data credits: Data by LLM Stats · Hugging Face official benchmark leaderboards; licence varies by benchmark · 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. · Steel.dev Agent Leaderboards, MIT (licence) · OpenEvals/leaderboard-data, MIT (licence)