CombEval — leaderboard
Metric: Avg. (self-reported). Source: benchmarklist.com. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 93.6 |
| 2 | Gemini 3 Flash (Preview) | 86.5 |
| 3 | GPT-5.1 | 74.1 |
| 4 | DeepSeek V3.2 | 71.4 |
| 5 | GPT-OSS-20B | 59.4 |
| 6 | Qwen 3 30B A3B 2507 (Thinking) | 39.4 |
| 7 | Llama 3.3 70B Instruct | 36.6 |
| 8 | Qwen2.5-Math-7B-Instruct | 26.3 |
| 9 | Llama 3 8B Instruct | 2 |
Interactive version: theaggregate.ai/benchmark?slug=combeval · How the rankings work · Data refreshed daily, snapshot 2026-07-22.