Table Question Answering on WikiTableQuestions
76.2AccuracyGPT-4o
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GPT-4oN=20002026.04 | 76.2 | 0.995 | 23.4 | 34.5 | 23.4 | 0.522 | |
| DeepSeek-V3N=20002026.04 | 69.3 | 0.985 | 29.2 | 40.4 | 29.2 | 0.543 | |
| Llama-3.3-70BN=20002026.04 | 69 | 0.985 | 29.6 | 40.1 | 30.2 | 0.523 | |
| GPT-4o-miniN=20002026.04 | 65.6 | 0.955 | 29.9 | 63.8 | 30 | 0.664 | |
| Gemini 2.5 FlashN=20002026.04 | 65.5 | 0.889 | 23.5 | 40 | 23.5 | 0.812 |