BrowseComp-V3 (Tool-Free): leaderboard
Metric: Success rate (%): share of questions answered correctly on 300 multimodal multi-hop browsing questions (Science 50, Technology 50, Society 75, Culture 75, Life 50) whose evidence is split across text and images on public web pages; tool-free: the question text and images only, no search or browsing, pass@1; higher is better. Source: arxiv.org. Saturation forecast: Around September 2028. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 12 | #78 |
| 2 | Seed 1.8 | 9 | #136 |
| 3 | O4 Mini | 7.33 | #172 |
| 4 | GPT-5.2 | 6 | #105 |
| 5 | Claude Sonnet 4.5 | 4 | #138 |
| 6 | Qwen 3 VL 235B A22B Instruct | 3.33 | #264 |
| 7 | MiMo-V2-Flash | 3 | #359 |
| 8 | GPT-4o | 2.67 | #333 |
| 9 | Qwen 3 VL 8B Instruct | 1 | #401 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=browsecomp-v3-tool-free · How It Works · Data refreshed daily, snapshot 2026-10-11.