Tennis Commentary Generation on TennisVL (test)
7.98BLEU-4TennisExpert
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| TennisExpertModel Category=Ours2026.03 | 7.98 | 31.54 | 29.16 | 43.71 | 15.79 | 19.29 | 16.84 | 16.73 | 19.4 | 88.05 | |
| Gemini-3-proModel Category=Proprietary, Setting=Zero-Shot2026.03 | 1.78 | 20.02 | 17.41 | 10.11 | 3.96 | 16.5 | 14.54 | 11.63 | 13.25 | 59.89 | |
| Gemini-3-flashModel Category=Proprietary, Setting=Zero-Shot2026.03 | 1.57 | 19.11 | 16.02 | 7.31 | 5.31 | 13.11 | 13.08 | 10.19 | 10.81 | 52.5 | |
| Qwen3.5-plusModel Category=Proprietary, Setting=Zero-Shot2026.03 | 1.51 | 21.74 | 16.61 | 5.44 | 2.7 | 15.21 | 14.01 | 9.49 | 11.14 | 52.55 | |
| GPT-5.2Model Category=Proprietary, Setting=Zero-Shot2026.03 | 1.32 | 18.39 | 15.95 | 7.52 | 3.73 | 16.02 | 14.26 | 10.82 | 11.76 | 56.59 | |
| Qwen3-VL-32BModel Category=Open-Weight, Setting=Zero-Shot2026.03 | 1.24 | 20.42 | 15.46 | 2.61 | 2.19 | 10.74 | 10.81 | 5.5 | 8.74 | 37.97 | |
| Claude-4.5-sonnetModel Category=Proprietary, Setting=Zero-Shot2026.03 | 1.22 | 19.99 | 14.2 | 2.09 | 3.06 | 15.76 | 12.9 | 9.19 | 9.84 | 50.75 | |
| Qwen3-VL-8BModel Category=Open-Weight, Setting=Zero-Shot2026.03 | 1.2 | 18.9 | 14.9 | 3.89 | 1.7 | 10.56 | 11.78 | 5.05 | 8.59 | 37.68 | |
| Qwen3.5-35BModel Category=Open-Weight, Setting=Zero-Shot2026.03 | 1.19 | 19.66 | 14.61 | 3.7 | 3.32 | 14.11 | 10.95 | 7.18 | 9.12 | 44.67 | |
| InternVL3.5-38BModel Category=Open-Weight, Setting=Zero-Shot2026.03 | 0.84 | 20.58 | 13.22 | 0.11 | 1.45 | 8.98 | 10.56 | 4.16 | 5.25 | 30.4 | |
| InternVL3.5-8BModel Category=Open-Weight, Setting=Zero-Shot2026.03 | 0.75 | 20.2 | 13.61 | 0.08 | 0.9 | 5.1 | 6.69 | 2.86 | 3.47 | 19.02 |