CodeAssay: leaderboard
Metric: Final hidden-test correctness (%; share of 185 audited Python tasks in ten software-engineering categories whose scored solution passes every hidden test; one generation from the description, signature and a public-test excerpt plus at most one repair guided by public-test failures; provider-default reasoning, temperature 0 where accepted, 16,384-token output limit; standard system prompt). Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 98.9 |
| 2 | Claude Sonnet 5 | 98.9 |
| 3 | Claude Haiku 4.5 | 96.8 |
| 4 | Claude Sonnet 4.5 | 95.1 |
| 5 | Gemini 2.5 Flash | 91.4 |
| 6 | GPT-4o | 87 |
| 7 | GPT-4o Mini | 77.3 |
Interactive version: theaggregate.ai/benchmark?slug=codeassay · How It Works · Data refreshed daily, snapshot 2026-09-29.