Skip to content

Benchmark

Open Agent Leaderboard BrowseCompPlus

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
5
Scores recorded
25
Sources
1
IQ weight
0.28

Score distribution

43.3%
Mean
46%
Median
10.6%
Std dev
30%
10th percentile
57%
90th percentile
21%
Min
21%61%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.28 for chat models, measured over 5 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 4.5agent: openai_solo, task: browsecompplus, route: openai/aws· Open Agent Leaderboard results data docs61%
2Claude Opus 4.5agent: smolagents_code, task: browsecompplus, route: openai/aws· Open Agent Leaderboard results data docs61%
3Gemini 3 Pro Previewagent: smolagents_code, task: browsecompplus, route: openai/gcp· Open Agent Leaderboard results data docs57%
4Kimi K2.5agent: claude_code, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs56%
5Claude Opus 4.5agent: claude_code, task: browsecompplus, route: openai/aws· Open Agent Leaderboard results data docs52.9%
6Gemini 3 Pro Previewagent: claude_code, task: browsecompplus, route: openai/gcp· Open Agent Leaderboard results data docs51%
7Claude Opus 4.5agent: tool_calling_with_shortlisting, task: browsecompplus, route: openai/aws· Open Agent Leaderboard results data docs49%
8Claude Opus 4.5agent: tool_calling, task: browsecompplus, route: openai/aws· Open Agent Leaderboard results data docs49%
9DeepSeek V3.2agent: claude_code, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs48%
10Gemini 3 Pro Previewagent: tool_calling_with_shortlisting, task: browsecompplus, route: openai/gcp· Open Agent Leaderboard results data docs48%
11Gemini 3 Pro Previewagent: tool_calling, task: browsecompplus, route: openai/gcp· Open Agent Leaderboard results data docs48%
12GPT 5.2agent: openai_solo, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs48%
13GPT 5.2agent: tool_calling_with_shortlisting, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs46%
14GPT 5.2agent: tool_calling, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs46%
15GPT 5.2agent: claude_code, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs43%
16DeepSeek V3.2agent: tool_calling_with_shortlisting, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs36%
17DeepSeek V3.2agent: tool_calling, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs36%
18Kimi K2.5agent: openai_solo, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs35%
19Kimi K2.5agent: tool_calling_with_shortlisting, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs34%
20Kimi K2.5agent: tool_calling, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs34%
21Gemini 3 Pro Previewagent: openai_solo, task: browsecompplus, route: openai/gcp· Open Agent Leaderboard results data docs33.3%
22Kimi K2.5agent: smolagents_code, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs33%
23DeepSeek V3.2agent: openai_solo, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs30%
24GPT 5.2agent: smolagents_code, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs26%
25DeepSeek V3.2agent: smolagents_code, task: browsecompplus, route: openai/azure· Open Agent Leaderboard results data docs21%

Data credits: Open Agent Leaderboard results (Exgentic), CDLA-Permissive-2.0; agreement text at https://cdla.dev/permissive-2-0/. Fractions shown as percentages. (licence)