PushupBench - R2: leaderboard
Agreement of predicted and annotated repetition counts on PushupBench (2026), 446 long-form exercise videos of 22 to 117 seconds, sampled at 5 frames per second: the coefficient of determination R2, at most 1, where 0 equals always predicting the mean count. Higher is better.
Metric: Coefficient of determination R2 between predicted and true repetition counts (at most 1; 0 matches predicting the mean, negative values mean constant or random output; predictions off by more than 50 excluded), on PushupBench's 446 long-form exercise clips (22-117 s, 11 creators in 10 countries; 27 clips accept several valid counts) sampled at 5 fps up to 112 frames (Claude models capped at 100 frames by the API), 360p frames, greedy decoding, 10 prompt templates; higher is better. Source: arxiv.org. Saturation forecast: Around March 2027. 20 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 0.82 | #175 |
| 2 | Gemini 3 Pro | 0.7 | #145 |
| 3 | Gemini 2.5 Pro | 0.63 | #276 |
| 4 | Gemini 2.5 Flash | 0.49 | #463 |
| 5 | Qwen 3 VL 32B Instruct | 0.36 | #606 |
| 6 | Qwen 3 VL 32B (Thinking) | 0.35 | #573 |
| 7 | Gemini 2.0 Flash | 0.25 | #723 |
| 8 | Qwen 3 VL 235B A22B Instruct | 0.17 | #532 |
| 9 | Claude Opus 4.5 | 0.09 | #160 |
| 10 | Qwen 3 VL 30B A3B (Thinking) | 0.08 | #719 |
| 11 | GPT-5 | 0.01 | #178 |
| 12 | Claude Sonnet 4.5 | 0 | #261 |
| 13 | Qwen 3 VL 30B A3B Instruct | -0.06 | #895 |
| 14 | Qwen 3 VL 8B Instruct | -0.11 | #1029 |
| 15 | Llama 4 Maverick | -0.15 | #1110 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=pushupbench-r2 · How It Works · Data refreshed daily, snapshot 2026-11-21.