SWE-bench Pro

SWE-bench Pro - refined benchmark with corrected problematic tasks in the public set. Internal agent scaffold; temp=1.0, top_p=0.95, 200K context window.

Category

Agentic Coding

Max Score

100.0

Score Type

percent

Active

Yes

Model Scores

Scores for SWE-bench Pro

Model Score Date Verified
Qwen3.6-35B-A3B
61.88%
16.04.2026 Verified
GPT 5.5
73.25%
11.06.2026 Unverified
Kimi K2.6
73.25%
11.06.2026 Unverified
MiniMax-M2.7
70.25%
11.06.2026 Unverified
Gemma4-31B-it
44.62%
16.04.2026 Unverified
Qwen3.5-35B-A3B
55.75%
16.04.2026 Unverified
Gemma4-26B-A4B
17.25%
16.04.2026 Unverified
Qwen3.6-27B
66.88%
22.04.2026 Verified
Qwen3.5-397B-A17B
63.62%
22.04.2026 Unverified
Claude 4.5 Opus
71.38%
22.04.2026 Unverified
Qwen3.5-27B
64.00%
16.04.2026 Unverified
Ornith-1.0-9B
53.62%
25.06.2026 Verified
Ornith-1.0-35B
63.00%
25.06.2026 Verified
Qwen3.8-27B
77.12%
18.08.2026 Verified
Qwen3.7-Plus
72.00%
18.08.2026 Unverified
Muse Glimmer-30B
64.00%
18.08.2026 Unverified
Opus4.6 Max
66.75%
18.08.2026 Unverified
Qwen3.8-2.4T-A95B
84.62%
18.08.2026 Verified
Fable-5
100.00%
18.08.2026 Unverified
GPT-5.6 Sol
80.75%
18.08.2026 Unverified
Claude Opus 4.8
86.50%
18.08.2026 Unverified
Muse-Glimmer-30B
64.00%
01.08.2026 Verified
DeepSeek V4 Pro
69.25%
11.06.2026 Unverified
Claude Opus 4.7
80.38%
11.06.2026 Unverified
MiniMax-M3
73.75%
11.06.2026 Verified
Gemini 3.1 Pro
67.75%
11.06.2026 Unverified
GLM-5.2
77.62%
17.06.2026 Verified
DeepSeek-V4-Pro
69.25%
17.06.2026 Unverified
GLM-5.1
73.00%
17.06.2026 Unverified
GPT-5.5
73.25%
17.06.2026 Unverified
Qwen3.7-Max
75.75%
17.06.2026 Unverified
Qwen3-Coder-Next
55.38%
03.02.2026 Verified
DeepSeek-V3.2
51.12%
03.02.2026 Verified
GLM-4.7
50.75%
03.02.2026 Verified
MiniMax M2.1
43.25%
03.02.2026 Verified
Claude-Opus-4.5
58.75%
03.02.2026 Verified
Claude-Sonnet-4.5
55.00%
03.02.2026 Verified
Kimi K2.5
50.00%
03.02.2026 Verified
Ornith-1.5-35B-A3B
74.50%
19.08.2026 Verified
Ornith-1.0-35B-A3B
63.00%
19.08.2026 Unverified
Gemma-4-31B
44.62%
19.08.2026 Unverified
GPT-5.4
72.12%
23.08.2026 Verified
Claude Opus 4.6
66.75%
23.08.2026 Verified
Opus-4.6
71.62%
26.06.2026 Verified
DeepSeek-V4-Pro
69.25%
26.06.2026 Verified
DeepSeek-V4-Flash
65.75%
26.06.2026 Verified
Gemma 3 27B IT
14.22%
27.02.2025 Verified
Qwen3.8-Flash-Next
78.12%
26.08.2026 Verified
DeepSeek-V4-Flash-0731
70.00%
26.08.2026 Unverified
Claude-Opus-4.6 (Max)
66.75%
26.08.2026 Unverified
Hy3
72.38%
28.08.2026 Verified
Hy4 preview
82.12%
28.08.2026 Verified
GLM-5.3
80.75%
28.08.2026 Verified
Kimi K3
79.12%
28.08.2026 Verified
Claude Opus 5
99.00%
28.08.2026 Verified
Spark-X2.5-4B
55.50%
— Verified
Spark-X2.5-1.7B
13.00%
— Verified
Qwen3.5-4B
36.75%
— Verified
Qwen3.5-2B
2.38%
— Verified
Gemma4-E4B
5.00%
— Verified
Qwen3.5-9B
42.25%
25.06.2026 Verified
Gemma4-12B
27.38%
25.06.2026 Verified
MiniCPM5-2B
18.00%
— Verified
LFM2.5-2.6B
0.75%
— Verified
Gemma4-E2B
0.00
— Verified
granite-4.2-3B
15.38%
— Verified
Nemotron-3-Nano-4B
0.12%
— Verified
LFM2.5-8B-A1B
0.50%
— Verified
GLM 5.1 Thinking
73.00%
11.06.2026 Unverified