GPQA Diamond (Epoch AI runs) leaderboard: who is #1

Name: GPQA Diamond (Epoch AI runs) rankings
Creator: Epoch AI
License: https://creativecommons.org/licenses/by/4.0/

Epoch AI · Accuracy, higher is better · 100 models · published Sep 2, 2026

Top 25 of 100 models, best variant per model as published

#	Model	Score
1	GPT-6 Astraas “gpt-6-astra_max”	95.8%
2	Gemini 3.8 Flashas “gemini-3.8-flash_high”	95.4%
3	Gemini 3.7 Flashas “gemini-3.7-flash_high”	94.8%
4	GPT-5.4 Proas “gpt-5.4-pro-2026-03-05_xhigh”	94.6%
5	Gemini 3.1 Pro Previewas “gemini-3.1-pro-preview_high”	94.4%
6	Gemini 3.6 Flashas “gemini-3.6-flash_high”	94.1%
7	GPT-5.5 Pre Releaseas “gpt-5.5-pre-release_xhigh”	94.0%
7	Grok 4.6as “grok-4.6_high”	94.0%
9	GPT-5.5 Pro Pre Releaseas “gpt-5.5-pro-pre-release_xhigh”	93.9%
10	Claude Opus 5as “claude-opus-5_max”	93.9%
11	GPT-5.6 Solas “gpt-5.6-sol_max”	93.5%
12	Grok 4.5as “grok-4.5_high”	93.4%
13	GPT-5.6 Terraas “gpt-5.6-terra_max”	93.3%
14	GPT-5.4as “gpt-5.4-2026-03-05_xhigh”	93.3%
15	Kimi K3as “kimi-k3_max”	93.1%
16	Gemini 3.5 Flashas “gemini-3.5-flash_high”	92.8%
17	Qwen3.8 Max (0902)as “qwen3.8-max_xhigh”	92.7%
18	Gemini 3 Proas “gemini-3-pro-preview”	92.6%
19	GLM-5.2as “glm-5.2_max”	91.9%
20	DeepSeek V4 Pro 0813as “deepseek-v4-pro-0813_max”	91.7%
21	GPT-5.6 Lunaas “gpt-5.6-luna_max”	91.6%
22	GPT-5.2as “gpt-5.2-2025-12-11_xhigh”	91.4%
23	Claude Opus 4.8as “claude-opus-4-8_max”	91.0%
23	DeepSeek V4 Flash Visionas “deepseek-v4-flash-0731_max”	91.0%
25	GLM-5.3as “glm-5.3_max”	90.9%