SWE-bench Pro (Anthropic Scaffold): leaderboard
Anthropic system-card run of SWE-bench Pro, measuring pass@1 on production software engineering issues using Anthropic scaffold settings.
Metric: Pass@1 (%). Source: www-cdn.anthropic.com. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Mythos 5 | 80.3 | #2 |
| 2 | Claude Fable 5 | 80 | #16 |
| 3 | Claude Mythos Preview | 77.8 | #4 |
| 4 | Claude Opus 4.8 | 69.2 | #37 |
| 5 | GPT-5.5 | 58.6 | #31 |
| 6 | Gemini 3.1 Pro (Preview) | 54.2 | #58 |
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-pro-anthropic-scaffold · How It Works · Data refreshed daily, snapshot 2026-10-20.