Benchgen

SWE-bench Pro — Results

RankModelScore
1fugu-ultra73.7
2qwen3-8-max67.7
3ornith-1-5-397b65.1
4gpt-5-6-sol64.6
5claude-opus-4-764.3
6gpt-5-6-terra63.4
7claude-sonnet-563.2
8gpt-5-6-luna62.7
9qwen3-8-flash-next62.5
10glm-5-262.1
11qwen3-8-27b61.7
12dots3-note-preview61
13ornith-1-5-35b-a3b59.6
14laguna-s-2-159.4
15fugu59
16gpt-5-558.6
17kimi-k2-658.6
18claude-sonnet-4-658.1
19seed-2-1-pro57.5
20mimo-v2-5-pro57.2
21seed-2-1-turbo57
22inkling54.3
23gemini-3-1-pro54.2
24muse-spark52.4
25muse-glimmer51.2

SWE-bench Pro — Leaderboard

1 phaseActive

SWE-bench Pro Leaderboard

No self-reported entries.