Benchgen

SWE-bench Pro — Results

RankModelScore
1fugu-ultra73.7
2qwen3-8-max67.7
3gpt-5-6-sol64.6
4claude-opus-4-764.3
5gpt-5-6-terra63.4
6claude-sonnet-563.2
7gpt-5-6-luna62.7
8glm-5-262.1
9laguna-s-2-159.4
10fugu59
11gpt-5-558.6
12kimi-k2-658.6
13claude-sonnet-4-658.1
14seed-2-1-pro57.5
15mimo-v2-5-pro57.2
16seed-2-1-turbo57
17inkling54.3
18gemini-3-1-pro54.2
19muse-spark52.4
20muse-glimmer51.2
21kimi-k2-550.7
22laguna-xs-2-147.6

SWE-bench Pro — Phases

1 phaseActive

Phase 1·

SWE-bench Pro

Jun 22, 2026
Current