FirmCORe - Role Direction: leaderboard
Metric: Macro-F1 (%). Source: arxiv.org. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Kimi K2.7 | 67.71 |
| 2 | Qwen 3.7 Plus | 67.04 |
| 3 | Gemma 4 26B | 66.52 |
| 4 | GLM-5.1 | 63.72 |
| 5 | Qwen 3.6 35B A3B | 63.38 |
| 6 | Gemma 3 12B | 61.02 |
| 7 | GLM-5.2 | 59.87 |
| 8 | Qwen3 Coder | 59.54 |
| 9 | Qwen 3.6 Max | 58.61 |
| 10 | Qwen 3.6 27B | 56.11 |
| 11 | Qwen 3 4B | 54.89 |
| 12 | Llama 3.1 8B | 52.25 |
| 13 | Qwen 3 8B | 52.1 |
| 14 | Llama 3.2 3B | 22.25 |
Interactive version: theaggregate.ai/benchmark?slug=firmcore-role-direction · How It Works · Data refreshed daily, snapshot 2026-09-20.