ClassEval: leaderboard
Class-level code generation benchmark evaluating class and function success under multiple prompting strategies.
Metric: Class-Level Pass@1 (%). Source: github.com. Source: fudanselab-classeval.github.io. Saturation forecast: Around January 2027. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 | 37.6 | #438 |
| 2 | GPT-3.5 | 29.6 | #818 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=classeval · How It Works · Data refreshed daily, snapshot 2026-10-20.