OpenAI GPT-5.5 Launch - Humanity's Last Exam (no tools): leaderboard
Metric: Score (%). Source: openai.com. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Mythos 5 | 59 |
| 2 | Claude Mythos Preview | 56.8 |
| 3 | Claude Opus 4.8 | 49.8 |
| 4 | Claude Opus 4.7 | 46.9 |
| 5 | Gemini 3.1 Pro (Preview) | 44.4 |
| 6 | GPT-5.5 Pro | 43.1 |
| 7 | GPT-5.4 Pro (xHigh) | 42.7 |
| 8 | GPT-5.5 | 41.4 |
| 9 | GPT-5.4 | 39.8 |
| 10 | GPT-5.2 Pro | 36.6 |
| 11 | GPT-5.2 | 34.5 |
Interactive version: theaggregate.ai/benchmark?slug=openai-gpt-5-5-launch-humanity-s-last-exam-no-tools · How It Works · Data refreshed daily, snapshot 2026-09-05.