Benchgen

DeepSWE — Results

RankModelScore
1gemini-4-argon77.9
2deepseek-v4-1-flash74.2
3deepseek-v4-1-flash74.2
4gpt-6-astra74.1
5mimo-v2-6-pro-rl71.9
6grok-4-771
7gpt-6-sol68.8
8mimo-v2-6-flash-rl67.9
9kimi-k367.5
10glm-5-366.9
11gpt-6-luna66.6
12grok-4-665.9
13nex-n2-5-max65.6
14hy4-preview64.3
15glm-5-3-flash63.4
16deepseek-v4-pro62.7
17muse-spark-1-259.3
18deepseek-v4-flash-vision-exp59.3
19qwen3-8-flash-next58.7
20qwen3-8-max56.6
21ornith-1-5-397b56
22nex-n2-5-pro55.8
23qwen3-8-27b42.2
24ornith-1-5-35b-a3b22
D

DeepSWE — Leaderboard

1 phaseActive

DeepSWE Leaderboard

No self-reported entries.