Document Arena

View overall rankings across AI models in document analysis and long-content reasoning.

Mar 8, 2026

39,396 votes

12 models

	Rank Spread
1	11	claude-opus-4-6 Anthropic · Proprietary	1526±12	4,040
2	23	claude-sonnet-4-6 Anthropic · Proprietary	1490±16	1,411
3	24	claude-opus-4-5-20251101 Anthropic · Proprietary	1473±11	5,819
4	36	gemini-3.1-pro-preview Google · Proprietary	1457±10	3,244
5	47	claude-sonnet-4-5-20250929 Anthropic · Proprietary	1450±11	6,033
6	47	gemini-3-pro Google · Proprietary	1447±9	8,434
7	511	claude-haiku-4-5-20251001 Anthropic · Proprietary	1429±12	5,313
8	710	gemini-2.5-pro Google · Proprietary	1428±9	6,291
9	712	gemini-3-flash Google · Proprietary	1422±9	7,293
10	712	gpt-5.2-high OpenAI · Proprietary	1412±9	5,379
11	812	gpt-5.1 OpenAI · Proprietary	1409±9	6,509
12	912	gpt-5.2 OpenAI · Proprietary	1408±9	8,104

Remove Style Control Leaderboard Plots

Fraction of Model A Wins for All Non-tied A vs. B Battles

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)

Confidence Intervals on Model Strength (via Bootstrapping)

Battle Count for Each Combination of Models (without Ties)