LongCat 2.5 Release - PaperBench: 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. Detailed table does not specify additional decoding controls; preserve effort as unspecified. Metric: Score (%).
Metric: Score (%). Source: longcat.ai. Saturation forecast: Around February 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 83.3 | #22 |
| 2 | GPT-5.5 | 80 | #33 |
| 3 | Claude Opus 5 | 79.1 | #16 |
| 4 | Kimi K3 | 77.5 | #35 |
| 5 | GLM-5.3 | 73.7 | #44 |
| 6 | Claude Opus 4.8 | 72.9 | #39 |
| 7 | LongCat 2.0 | 45.2 | #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-paperbench · How It Works · Data refreshed daily, snapshot 2026-11-05.