Share your thoughts, 1 month free Claude Pro on usSee more

Coding Agent Capacity on SWE-Bench Multi

74Score

Claude-Opus-4.5

Updated 23d ago

Evaluation Results

Method	Links
Claude-Opus-4.5 2026.06		74
Gemini-3-pro High 2026.06		72.7
Seed2.0 Pro 2026.06		71.7
GPT-5.2 High 2026.06		68.8
Claude-Sonnet-4.5 2026.06		64.1
Gemini-3-Flash High 2026.06		59
Claude-Opus-4.5 2026.06		52.8
Gemini-3-pro High 2026.06		50.2
GPT-5-mini High 2026.06		49.3
GPT-5.2 High 2026.06		47.7
Claude-Sonnet-4.5 2026.06		47.7
Seed2.0 Pro 2026.06		45.2
Seed2.0 Lite 2026.06		41.1