Exgentic Open Agent - BrowseComp+: leaderboard
Exgentic Open Agent BrowseComp+ evaluates deep-research agents on hard web-browsing questions.
Metric: Success Rate (%). Source: www.exgentic.ai. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | OpenAI Solo + claude-opus-4.5 | 61 | |
| 2 | Smolagent + claude-opus-4.5 | 61 | |
| 3 | Smolagent + gemini-3-pro | 57 | |
| 4 | Claude Code + kimi-k2.5 | 56 | |
| 5 | Claude Code + claude-opus-4.5 | 52.94 | |
| 6 | Claude Code + gemini-3-pro | 51 | |
| 7 | React + claude-opus-4.5 | 49 | |
| 8 | React + Shortlisting + claude-opus-4.5 | 49 | |
| 9 | Claude Code + deepseek-v3.2 | 48 | |
| 10 | OpenAI Solo + gpt-5.2 | 48 |
Interactive version: theaggregate.ai/benchmark?slug=exgentic-open-agent-browsecomp-plus · How It Works · Data refreshed daily, snapshot 2026-10-20.