mmlu Pro Arcee
6 models evaluated
|
| 1 | Claude Opus 4.6 | Anthropic | Closed | 89.1% | |
| 2 | Kimi K2.5 | Moonshot AI | Open | 87.1% | |
| 3 | GLM-5 | Z.AI | Open | 85.8% | |
| 4 | Trinity-Large-Thinking | Arcee AI | Open | 83.4% | |
| 5 | MiniMax M2.7 | MiniMax | Open | 80.8% | |
| 6 | Trinity-Large-Preview | Arcee AI | Open | 75.2% | |