Humanity's Last Exam leaderboard: who is #1

Name: Humanity's Last Exam rankings
Creator: Scale AI and CAIS
License: https://creativecommons.org/licenses/by/4.0/

Scale AI and CAIS · Accuracy, higher is better · 42 models · published Sep 24, 2026

Top 25 of 42 models, best variant per model as published

#	Model	Score
1	GPT-6 Astraas “GPT 6 Astra”	54.8%
2	Claude Fable 5.1as “Fable 5.1 (xhigh)”	46.5%
3	Gemini 3.1 Pro Previewas “gemini-3.1-pro-preview (thinking high)”	46.4%
4	Gemini 3.8 FlashGoogle	44.5%
5	GPT-5.4 Proas “gpt-5.4-pro-2026-03-05”	44.3%
6	Muse SparkMeta	40.6%
7	Gemini 3 Proas “gemini-3-pro-preview”	37.5%
8	GPT-5.4as “gpt-5.4-2026-03-05 (xhigh thinking)”	36.2%
9	Claude Opus 4.7as “claude-opus-4-7”	36.2%
10	Claude Opus 4.6as “claude-opus-4-6-thinking-max”	34.4%
11	GPT-5 Proas “gpt-5-pro-2025-10-06”	31.6%
12	GPT-5.2as “gpt-5.2-2025-12-11”	27.8%
13	GPT-5as “gpt-5-2025-08-07”	25.3%
14	Claude Opus 4.5as “claude-opus-4-5-20251101-thinking”	25.2%
15	Kimi K2.5as “kimi-k2.5”	24.4%
16	GPT-5.1as “gpt-5.1-thinking”	23.7%
17	Gemini 2.5 Pro (Jun 2025)as “gemini-2.5-pro-preview-06-05”	21.6%
18	o3as “o3 (high) (April 2025)”	20.3%
19	GPT-5 minias “gpt-5-mini-2025-08-07”	19.4%
20	o4-minias “o4-mini (high) (April 2025)”	18.1%
21	Claude Sonnet 4.5as “claude-sonnet-4-5-20250929-thinking”	13.7%
22	Gemini 2.5 Flash (Sep 2025)as “Gemini 2.5 Flash (April 2025)”	12.1%
23	Claude Opus 4.1as “claude-opus-4-1-20250805-thinking”	11.5%
24	Claude Opus 4as “Claude Opus 4 (Thinking)”	10.7%
25	Gemini 3.1 Flash-Liteas “gemini-3.1-flash-lite-preview”	8.6%