# Tennis Commentary Generation on TennisVL (test) Source: [Tennis Commentary Generation on TennisVL (test)]() Collection index: [llms.txt]() Data updated: 2026-03-17T08:00:09.807Z Evaluation of models generating commentary for tennis videos using the TennisVL dataset (test split), assessed using standard NLP metrics (BLEU, ROUGE) alongside custom subjective evaluations (Coherence, Excitement, Professionalism). A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Tennis Commentary Generation | | Dataset (including any stated split) | TennisVL (test) | | Metric | BLEU-4 | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 11 of 11 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [BLEU-4](): Higher is better; unit: SCALAR. - [METEOR](): Higher is better; unit: SCALAR. - [ROUGE-L](): Higher is better; unit: SCALAR. - [CIDEr](): Higher is better; unit: SCALAR. - [Accuracy](): Higher is better; unit: PERCENT. - [Coherence](): Higher is better; unit: SCALAR. - [Excitement](): Higher is better; unit: SCALAR. - [Professionalism](): Higher is better; unit: SCALAR. - [Pacing](): Higher is better; unit: SCALAR. - [Total Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | BLEU-4 | Result date | Report paper | | --- | --- | --- | --- | | TennisExpert | 7.98 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Gemini-3-pro | 1.78 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Gemini-3-flash | 1.57 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Qwen3.5-plus | 1.51 | Not recorded | 69b8e00a0636e5bf4afbea01 | | GPT-5.2 | 1.32 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Qwen3-VL-32B | 1.24 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Claude-4.5-sonnet | 1.22 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Qwen3-VL-8B | 1.2 | Not recorded | 69b8e00a0636e5bf4afbea01 | | Qwen3.5-35B | 1.19 | Not recorded | 69b8e00a0636e5bf4afbea01 | | InternVL3.5-38B | 0.84 | Not recorded | 69b8e00a0636e5bf4afbea01 | | InternVL3.5-8B | 0.75 | Not recorded | 69b8e00a0636e5bf4afbea01 | ### TennisExpert BLEU-4: 7.98. Result ID: 69b9093d47f34ce924a158dd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Ours"}. This record does not assert independent reproduction. ### Gemini-3-pro BLEU-4: 1.78. Result ID: 69b9093d47f34ce924a158dc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Proprietary","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### Gemini-3-flash BLEU-4: 1.57. Result ID: 69b9093d47f34ce924a158da. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Proprietary","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### Qwen3.5-plus BLEU-4: 1.51. Result ID: 69b9093d47f34ce924a158d9. [Reporting paper]() Recorded evaluation configuration: {"Model Category":"Proprietary","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### GPT-5.2 BLEU-4: 1.32. Result ID: 69b9093d47f34ce924a158db. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Proprietary","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### Qwen3-VL-32B BLEU-4: 1.24. Result ID: 69b9093d47f34ce924a158d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Open-Weight","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### Claude-4.5-sonnet BLEU-4: 1.22. Result ID: 69b9093d47f34ce924a158d8. [Reporting paper]() Recorded evaluation configuration: {"Model Category":"Proprietary","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### Qwen3-VL-8B BLEU-4: 1.2. Result ID: 69b9093d47f34ce924a158d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Open-Weight","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### Qwen3.5-35B BLEU-4: 1.19. Result ID: 69b9093d47f34ce924a158d7. [Reporting paper]() Recorded evaluation configuration: {"Model Category":"Open-Weight","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### InternVL3.5-38B BLEU-4: 0.84. Result ID: 69b9093d47f34ce924a158d4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Open-Weight","Setting":"Zero-Shot"}. This record does not assert independent reproduction. ### InternVL3.5-8B BLEU-4: 0.75. Result ID: 69b9093d47f34ce924a158d3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Open-Weight","Setting":"Zero-Shot"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.