Back to Benchmarks
🤖

BrowseComp

Browse Competition

Web research agent benchmark

🤖 Agentic / Tool Use
Best: 84%
8 Models Evaluated
Champion

Claude Opus 4.6

Anthropic

84%

Best Score

Model Rankings

All models ranked by score on BrowseComp. Scores are relative to the best score of 84%.

#ModelProviderScorevs Best
1Claude Opus 4.6
Elo 1503
Anthropic
84%
Best
2DeepSeek V4 Pro (Max)DeepSeek
83.4%
-0.6%
3GPT-5.5
Elo 1478
OpenAI
82.7%
-1.3%
4GPT-5.4
Elo 1463
OpenAI
82.7%
-1.3%
5Qwen 3.5
Elo 1450
Alibaba
78.6%
-5.4%
6GLM-5
Elo 1454
Zhipu AI
75.9%
-8.1%
7Claude Sonnet 4.6
Elo 1460
Anthropic
74.7%
-9.3%
8MiMo-V2-Flash
Elo 1393
Xiaomi
58.3%
-25.7%

Score Distribution

Claude Opus 4.6
84%
DeepSeek V4 Pro (Max)
83.4%
GPT-5.5
82.7%
GPT-5.4
82.7%
Qwen 3.5
78.6%
GLM-5
75.9%
Claude Sonnet 4.6
74.7%
MiMo-V2-Flash
58.3%

Run This Benchmark

Learn how to evaluate your own models on BrowseComp.

BrowseComp is a standardized benchmark for evaluating AI model capabilities.

To run this benchmark on your own models, visit the official benchmark page or leaderboard for detailed instructions, code, and evaluation scripts.

Data compiled from official model cards, public leaderboards, and independent evaluations.