Kaggle VideoQA: leaderboard
Google DeepMind's VideoQA leaderboard on Kaggle: question answering over video inputs, scored as accuracy and shown as 0-100 (higher is better). The roster is small and Gemini-heavy (9 of 13 models on 2026-09-13), so treat rankings outside that family with care.
Metric: Accuracy (%). Source: www.kaggle.com. Saturation forecast: Not forecast. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.7 Flash | 89.11 | #24 |
| 2 | Gemini 3.8 Flash | 87.13 | #30 |
| 3 | Gemini 4 Argon | 87 | #9 |
| 4 | Gemini 3.6 Flash | 82.18 | #57 |
| 5 | Gemini 3.5 Flash | 80.2 | #66 |
| 6 | Gemini 3 Flash (Preview) | 79.21 | #90 |
| 7 | Gemini 2.5 Pro | 78.95 | #159 |
| 8 | Gemini 3.1 Pro (Preview) | 78.22 | #64 |
| 9 | Gemini 3.1 Flash Lite (Preview) | 78.22 | #201 |
| 10 | GPT-6 Astra | 75.25 | #6 |
| 11 | Gemini 3.5 Flash Lite | 75.25 | #174 |
| 12 | Claude Opus 4.8 | 71.29 | #39 |
| 13 | Claude Haiku 5.5 | 68.32 | #43 |
| 14 | Claude Sonnet 4.6 | 67.33 | #94 |
| 15 | GPT-5.6 Terra | 67.33 | #40 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=kaggle-videoqa · How It Works · Data refreshed daily, snapshot 2026-11-05.