Multiple-choice Question Answering on KNIGHT Math L3
89.4AccuracyHuman
Evaluation Results
| Method | Links | |
|---|---|---|
| Humann=2002026.02 | 89.4 | |
| GPT-4o2026.02 | 86.41 | |
| Llama3-70B-InstructSize=70B, Type=Instruct2026.02 | 84.67 | |
| Mistral Large2026.02 | 84.1 | |
| Claude 3 Haiku2026.02 | 81.6 | |
| Qwen1.5 (1.8B)Size=1.8B2026.02 | 71.43 | |
| Gemma (2B)Size=2B2026.02 | 40.45 |