Search Arena

View rankings across LLMs with web search for real-time information, external knowledge, and grounded citations Learn more about Search Arena at our blog

Jan 29, 2026

183,043 votes

19 models

/

/

Style Control

	Rank Spread
1	13	gemini-3-flash-grounding Google · Proprietary	1224±6	11,062
2	13	gemini-3-pro-grounding Google · Proprietary	1219±6	18,839
3	14	gpt-5.2-search OpenAI · Proprietary	1218±7	12,157
4	34	gpt-5.1-search OpenAI · Proprietary	1207±6	14,152
5	57	gpt-5.2-search-non-reasoning OpenAI · Proprietary	1189±8	5,510
6	57	grok-4-1-fast-search xAI · Proprietary	1185±6	14,111
7	58	claude-opus-4-5-search Anthropic · Proprietary	1179±10	4,293
8	78	grok-4-fast-search xAI · Proprietary	1170±5	31,388
9	914	o3-search OpenAI · Proprietary	1144±5	21,056
10	914	gemini-2.5-pro-grounding Google · Proprietary	1143±4	36,828
11	916	ppl-sonar-reasoning-pro-high Perplexity · Proprietary	1143±5	29,825
12	916	grok-4-search xAI · Proprietary	1142±5	19,628
13	917	claude-sonnet-4-5-search Anthropic · Proprietary	1142±10	4,348
14	917	claude-opus-4-1-search Anthropic · Proprietary	1139±4	36,199
15	1117	gpt-5-search OpenAI · Proprietary	1133±5	21,212
16	1117	ppl-sonar-pro-high Perplexity · Proprietary	1133±5	29,379
17	1317	claude-opus-4-search Anthropic · Proprietary	1132±5	32,002
18	1819	diffbot-small-xl Diffbot · Apache 2.0	1024±8	6,473
19	1819	api-gpt-4o-search OpenAI · Proprietary	1008±11	3,399

Default Leaderboard Plots

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

Confidence Intervals on Model Strength (via Bootstrapping)

Battle Count for Each Combination of Models (without Ties)

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