Named-entity recognition on Side-effect entity extraction gold-standard dataset
91.2Side Effect ScoreDeepseek-V3
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Deepseek-V32026.06 | 91.2 | 64.1 | 42.9 | 33.3 | |
| Claude-Sonnet-42026.06 | 87.9 | 73.2 | 35 | 50 | |
| GPT-4.1-mini2026.06 | 85.8 | 57.1 | 38.9 | 71.4 | |
| Gemini-2.5-Flash2026.06 | 84.6 | 74.2 | 33.3 | 75 | |
| GPT-5-mini2026.06 | 83.7 | 69 | 47.6 | 66.7 | |
| Qwen3-235b-A22b2026.06 | 80 | 51.4 | 35.3 | 14.3 | |
| GPT-4.1-nano2026.06 | 76.5 | 63.6 | 30 | 66.7 | |
| GPT-4o-mini2026.06 | 74.9 | 39.5 | 18.8 | 12.5 | |
| GPT-5-nano2026.06 | 71.5 | 41.7 | 46.2 | 20 |