LongCat 2.5 Release - BrowseComp: leaderboard
LongCat-2.5 final release detailed native result table, November 2026. External:true cells are explicitly cited other-evaluator results and excluded from this component; self-run numeric cells retained. No Preview/final or release-generation aliases. Detailed table used instead of inconsistent rounded headline charts. BrowseComp: Custom lightweight harness with Verify and Discard-All modules (Discard-All threshold 256K), capped at 1,024 exploration rounds; only the `search` and `browse_with_goal` tools are available. Metric: Score (%).
Metric: Score (%). Source: longcat.ai. Saturation forecast: Estimated already saturated. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4.1 Flash | 91.5 | #37 |
| 2 | GLM-5.3 | 81 | #44 |
| 3 | LongCat 2.0 | 79.9 | #220 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=longcat-2-5-release-browsecomp · How It Works · Data refreshed daily, snapshot 2026-11-05.