BrowseComp Long Context 256k: leaderboard
BrowseComp is a benchmark for measuring the ability of agents to browse the web, comprising 1,266 questions that require persistently navigating the internet in search of hard-to-find, entangled information. Despite the difficulty of the questions, BrowseComp is simple and easy-to-use, as predicted answers are short and easily verifiable against reference answers. The benchmark focuses on questions where answers are obscure, time-invariant, and well-supported by evidence scattered across the open web.
Metric: Score (%). Source: openai.com. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 88.8 | #98 |
| 2 | GPT-5 Mini | 86 | #186 |
| 3 | GPT-4.1 Mini | 81.6 | #355 |
| 4 | GPT-4.1 | 75.5 | #248 |
| 5 | GPT-5 Nano | 68.4 | #425 |
| 6 | GPT-4.1 Nano | 19.1 | #731 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=browsecomp-long-context-256k · How It Works · Data refreshed daily, snapshot 2026-10-20.