AGC-Bench - rpgbench: leaderboard
Metric: Dataset z-score. Source: huggingface.co. 82 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K2.5 | 1.74 |
| 2 | GPT-5.4 | 1.6 |
| 3 | Kimi K2 0905 | 1.54 |
| 4 | Kimi K2.6 | 1.51 |
| 5 | GPT-5.5 | 1.34 |
| 6 | Qwen 3 Next 80B A3B Instruct | 1.34 |
| 7 | GPT-5.1 | 1.25 |
| 8 | Claude Sonnet 4.6 | 1.25 |
| 9 | DeepSeek V3.2 | 1.14 |
| 10 | GPT-5.4 Mini | 1.13 |
| 11 | Nemotron 3 Super | 1.09 |
| 12 | DeepSeek V3.2 Exp | 1.07 |
| 13 | Claude Opus 4.7 | 1.06 |
| 14 | Mistral Medium 3.1 | 1.02 |
| 15 | DeepSeek V3.1 Terminus | 1.01 |
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-rpgbench · How It Works · Data refreshed daily, snapshot 2026-09-05.