SWE-bench Multimodal (Anthropic Internal Harness): leaderboard

Anthropic internal multimodal SWE-bench harness, measuring software issue resolution that requires visual or multimodal context.

Metric: Resolved (%). Source: www-cdn.anthropic.com. Saturation forecast: Around April 2028. 2 models tracked.

Top models

#ModelScoreOverall rank
1Claude Opus 4.838.4#37
2Claude Opus 4.734.5#49

No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=swe-bench-multimodal-anthropic-internal-harness · How It Works · Data refreshed daily, snapshot 2026-10-20.