HAL SWE-bench Verified Mini: leaderboard
HAL SWE-bench Verified Mini evaluates software issue resolution on a compact SWE-bench Verified subset.
Metric: Score (%). Source: www.swebench.com. Saturation forecast: Around January 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 (High) | 72 | #142 (Claude Sonnet 4.5) |
| 2 | Claude Sonnet 4.5 | 68 | #142 |
| 3 | Claude Opus 4.1 (20250805) | 61 | #132 |
| 4 | O4 Mini (Low) | 54 | #180 (O4 Mini) |
| 5 | Claude 3.7 Sonnet (20250219) | 50 | #201 |
| 6 | Claude Opus 4 (20250514) | 50 | #154 |
| 7 | O4 Mini (High) | 50 | #180 (O4 Mini) |
| 8 | GPT-5 (Medium) | 46 | #98 (GPT-5) |
| 9 | O3 (Medium) | 46 | #124 (O3) |
| 10 | GPT-4.1 (2025-04-14) | 44 | #250 |
| 11 | Claude Haiku 4.5 (High) | 44 | #280 (Claude Haiku 4.5) |
| 12 | Claude Haiku 4.5 (20251001) | 24 | #261 |
| 13 | Gemini 2.0 Flash (001) | 24 | #360 |
| 14 | DeepSeek R1 | 6 | #251 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=hal-swe-bench-verified-mini · How It Works · Data refreshed daily, snapshot 2026-10-20.