HiL-Bench Original Paper - SWE - Blocker Recall: leaderboard
Metric: Blocker Recall (%). Source: static.scale.com. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 36.2 |
| 2 | Claude Opus 4.6 | 34.6 |
| 3 | GPT-5.4 | 27 |
| 4 | GPT-5.3 Codex | 23.5 |
Interactive version: theaggregate.ai/benchmark?slug=hil-bench-original-paper-swe-blocker-recall · How It Works · Data refreshed daily, snapshot 2026-09-19.