Multiple-choice Question Answering on KNIGHT History L3
89.2AccuracyHuman
Evaluation Results
| Method | Links | |
|---|---|---|
| Humann=2002026.02 | 89.2 | |
| GPT-4o2026.02 | 86.39 | |
| Llama3-70B-InstructSize=70B, Type=Instruct2026.02 | 85.15 | |
| Mistral Large2026.02 | 84.16 | |
| Claude 3 Haiku2026.02 | 81.64 | |
| Qwen1.5 (1.8B)Size=1.8B2026.02 | 72.45 | |
| Gemma (2B)Size=2B2026.02 | 30.73 |