HLE-Verified (BenchLM): leaderboard
Expert questions from the verified and revised 1,811-item Humanity’s Last Exam set; full-set accuracy reported in BenchLM’s attributed comparison table. These are published results, not BenchLM ability estimates.
Metric: Accuracy (%). Source: benchlm.ai. Saturation forecast: Around May 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.8 Flash | 54.9 | #30 |
| 2 | GPT-5.6 Sol | 54.5 | #22 |
| 3 | Claude Opus 5 | 54.4 | #16 |
| 4 | Gemini 3.7 Flash | 53.6 | #24 |
| 5 | GPT-5.6 Terra | 51.1 | #40 |
| 6 | Claude Sonnet 5 | 31 | #65 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=hle-verified-benchlm · How It Works · Data refreshed daily, snapshot 2026-11-05.