Multiple-choice Question Answering on CBT-Bench Task 1 Q&A
96Q&A AccuracyDeepSeek-R1-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-R1-8BParameters=8B2026.07 | 96 | |
| GPT-OSSParameters=20B2026.07 | 89.86 | |
| Gemma3Parameters=12B2026.07 | 80.45 | |
| MistralParameters=7B2026.07 | 74.55 |