Leaderboard

Rank	Diff	Model	Score	95% CI	Votes	License
1	—	grok-4-fast xAI	1174		32	Proprietary
2	—	gemini-3-pro-preview Google	1095		28	Proprietary
3	—	deepseek-v3.1-terminus DeepSeek	1091		37	Open Source
4	—	claude-opus-4.1 Anthropic	1056		33	Proprietary
5	—	claude-sonnet-4.6 Anthropic	1054		21	Proprietary
6	—	gpt-4.1 OpenAI	1035		34	Proprietary
7	—	gemini-3.1-pro-preview Google	1026		11	Proprietary
8	—	grok-4.1-fast xAI	1022		28	Proprietary
9	—	qwen3-vl-235b-a22b-instruct Qwen	991		24	Open Source
10	—	grok-4 xAI	989		33	Proprietary
11	—	gpt-5.1 OpenAI	986		25	Proprietary
12	—	kimi-k2-0905 Moonshot AI	951		34	Open Source
13	—	gpt-5.2 OpenAI	938		22	Proprietary
14	—	deepseek-chat-v3.1 DeepSeek	919		22	Open Source
15	—	gpt-5-chat OpenAI	917		34	Proprietary
16	—	o3 OpenAI	913		24	Proprietary
17	—	glm-4.6 Zhipu AI	906		26	Open Source
18	—	deepseek-r1-0528 DeepSeek	904		37	Open Source
19	—	qwen3-max Qwen	897		36	Proprietary
20	—	claude-opus-4 Anthropic	890		33	Proprietary
21	—	qwen3-235b-a22b-2507 Qwen	867		29	Open Source
22	—	gpt-4o OpenAI	837		12	Proprietary
23	—	gemini-2.5-pro Google	834		30	Proprietary
24	—	claude-sonnet-4.5 Anthropic	833		37	Proprietary
25	—	kimi-k2 Moonshot AI	824		26	Open Source

Models are ranked by Bradley-Terry score on an Elo scale. 95% bootstrap confidence intervals show statistical uncertainty; once significance-aware ranking is enabled, models with overlapping intervals will share a tied rank.