Multiple-choice Question Answering on KNIGHT Math L1
97.4AccuracyHuman
Evaluation Results
| Method | Links | |
|---|---|---|
| Humann=2002026.02 | 97.4 | |
| GPT-4o2026.02 | 95.3 | |
| Mistral Large2026.02 | 95.07 | |
| Llama3-70B-InstructSize=70B, Type=Instruct2026.02 | 95.03 | |
| Claude 3 Haiku2026.02 | 93.97 | |
| Qwen1.5 (1.8B)Size=1.8B2026.02 | 75.84 | |
| Gemma (2B)Size=2B2026.02 | 43.62 |