OpenAI GPT-5.5 Launch - Humanity's Last Exam (no tools): leaderboard

Metric: Score (%). Source: openai.com. 6 models tracked.

Top models

#ModelScore
1Claude Mythos 559
2Claude Mythos Preview56.8
3Claude Opus 4.849.8
4Claude Opus 4.746.9
5Gemini 3.1 Pro (Preview)44.4
6GPT-5.5 Pro43.1
7GPT-5.4 Pro (xHigh)42.7
8GPT-5.541.4
9GPT-5.439.8
10GPT-5.2 Pro36.6
11GPT-5.234.5

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.