AnyGroundBench - Sports TVG: leaderboard
Metric: tIoU@0.3 (%; share of queries whose temporal IoU is at least 0.3) on the 163 sports-domain test queries (MultiSports and newly captured American football videos), temporal video grounding (the queried event's start and end), zero-shot, default inference settings (Qwen3.5 with thinking disabled); higher is better. Source: arxiv.org. Saturation forecast: Around November 2027. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 37.4 |
| 2 | Gemini 2.5 Flash | 31.9 |
| 3 | Gemini 3 Flash | 29.4 |
| 4 | Gemini 3.1 Pro (Preview) | 26.3 |
| 5 | GPT-5.1 | 25.7 |
| 6 | GPT-4o | 17.1 |
| 7 | Qwen 3.5 9B (Non-reasoning) | 14.7 |
| 8 | Qwen 3.5 4B (Non-reasoning) | 12.8 |
| 9 | InternVL3-14B | 7.36 |
| 10 | InternVL3-8B | 6.79 |
| 11 | InternVL3.5-8B | 6.13 |
Interactive version: theaggregate.ai/benchmark?slug=anygroundbench-sports-tvg · How It Works · Data refreshed daily, snapshot 2026-09-29.