Skip to content

Benchmark

Multi-SWE-bench C Verified

Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.

Models measured
11
Scores recorded
31
Sources
1
IQ weight
1.64

Score distribution

3.6%
Mean
3.1%
Median
2.9%
Std dev
0.781%
10th percentile
8.6%
90th percentile
0%
Min
0%10.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.64 for chat models, measured over 11 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Sonnet 4RepoRepair, 20251226_RepoRepair_Claude-4-Sonnet· Multi-SWE-bench Experiments data docs10.9%
2Gemini 2.5 ProMagentLess, 20250426_MagentLess_Gemini-2.5-Pro· Multi-SWE-bench Experiments data docs9.4%
3Gemini 2.5 ProMSWE-agent, 20250426_MSWE-agent_Gemini-2.5-Pro· Multi-SWE-bench Experiments data docs9.4%
4Claude Sonnet 3.7MopenHands, 20250329_MopenHands_Claude-3.7-Sonnet· Multi-SWE-bench Experiments data docs8.6%
5Claude Sonnet 3.7MSWE-agent, 20250329_MSWE-agent_Claude-3.7-Sonnet· Multi-SWE-bench Experiments data docs8.6%
6Claude Sonnet 3.7MagentLess, 20250329_MagentLess_Claude-3.7-Sonnet· Multi-SWE-bench Experiments data docs6.3%
7Gemini 2.5 ProMopenHands, 20250426_MopenHands_Gemini-2.5-Pro· Multi-SWE-bench Experiments data docs5.5%
8Claude 3.5 SonnetMSWE-agent, 20250329_MSWE-agent_Claude-3.5-Sonnet(Oct)· Multi-SWE-bench Experiments data docs4.7%
9Llama 4 MaverickMagentLess, 20250426_MagentLess_Llama-4-Maverick· Multi-SWE-bench Experiments data docs3.9%
10o1MSWE-agent, 20250329_MSWE-agent_OpenAI-o1· Multi-SWE-bench Experiments data docs3.9%
11Claude 3.5 SonnetMopenHands, 20250329_MopenHands_Claude-3.5-Sonnet(Oct)· Multi-SWE-bench Experiments data docs3.1%
12DeepSeek V3MagentLess, 20250329_MagentLess_DeepSeek-V3· Multi-SWE-bench Experiments data docs3.1%
13DeepSeek V3MopenHands, 20250329_MopenHands_DeepSeek-V3· Multi-SWE-bench Experiments data docs3.1%
14Llama 4 MaverickMopenHands, 20250426_MopenHands_Llama-4-Maverick· Multi-SWE-bench Experiments data docs3.1%
15o1MagentLess, 20250329_MagentLess_OpenAI-o1· Multi-SWE-bench Experiments data docs3.1%
16o1MopenHands, 20250329_MopenHands_OpenAI-o1· Multi-SWE-bench Experiments data docs3.1%
17DeepSeek R1MopenHands, 20250329_MopenHands_DeepSeek-R1· Multi-SWE-bench Experiments data docs2.3%
18DeepSeek V3MSWE-agent, 20250329_MSWE-agent_DeepSeek-V3· Multi-SWE-bench Experiments data docs2.3%
19o3 MiniMagentLess, 20250329_MagentLess_OpenAI-o3-mini-high· Multi-SWE-bench Experiments data docs2.3%
20o3 MiniMSWE-agent, 20250329_MSWE-agent_OpenAI-o3-mini-high· Multi-SWE-bench Experiments data docs2.3%
21GPT 4oMagentLess, 20250329_MagentLess_GPT-4o-1120· Multi-SWE-bench Experiments data docs1.6%
22GPT 4oMSWE-agent, 20250329_MSWE-agent_GPT-4o-1120· Multi-SWE-bench Experiments data docs1.6%
23o3 MiniMopenHands, 20250329_MopenHands_OpenAI-o3-mini-high· Multi-SWE-bench Experiments data docs1.6%
24Qwen2.5 72B InstructMagentLess, 20250329_MagentLess_Qwen2.5-72B-Instruct· Multi-SWE-bench Experiments data docs1.6%
25Qwen2.5 72B InstructMSWE-agent, 20250329_MSWE-agent_Qwen2.5-72B-Instruct· Multi-SWE-bench Experiments data docs1.6%
26Claude 3.5 SonnetMagentLess, 20250329_MagentLess_Claude-3.5-Sonnet(Oct)· Multi-SWE-bench Experiments data docs0.781%
27DeepSeek R1MagentLess, 20250329_MagentLess_DeepSeek-R1· Multi-SWE-bench Experiments data docs0.781%
28DeepSeek R1MSWE-agent, 20250329_MSWE-agent_DeepSeek-R1· Multi-SWE-bench Experiments data docs0.781%
29Qwen2.5 72B InstructMopenHands, 20250329_MopenHands_Qwen2.5-72B-Instruct· Multi-SWE-bench Experiments data docs0.781%
30GPT 4oMopenHands, 20250329_MopenHands_GPT-4o-1120· Multi-SWE-bench Experiments data docs0%
31Llama 4 MaverickMSWE-agent, 20250426_MSWE-agent_Llama-4-Maverick· Multi-SWE-bench Experiments data docs0%

Data credits: Multi-SWE-bench experiments (multi-swe-bench/experiments), Apache-2.0; resolved-rate fractions converted to percentages (licence)