JudgeBench Coding — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 52 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 97.62 |
| 2 | GPT-5 Mini | 97.62 |
| 3 | GPT-5.5 | 97.62 |
| 4 | Gemma 4 31B (IT) | 97.62 |
| 5 | Kimi K2.5 | 97.62 |
| 6 | Kimi K2.6 | 97.62 |
| 7 | DeepSeek R1 0528 | 97.62 |
| 8 | Qwen 3.5 35B A3B | 97.62 |
| 9 | Step 3.5 Flash | 97.62 |
| 10 | DeepSeek V3.2 Speciale | 97.62 |
| 11 | GLM-4.7 FP8 | 97.62 |
| 12 | GLM-5 FP8 | 97.62 |
| 13 | GLM-5.1 FP8 | 97.62 |
| 14 | GPT-OSS-20B | 96.43 |
| 15 | DeepSeek-V4-Flash-FP8 | 96.43 |
Interactive version: theaggregate.ai/benchmark?slug=judgebench-coding · How the rankings work · Data refreshed daily, snapshot 2026-07-22.