BARISTA - Activity Recognition: leaderboard
Metric: Accuracy (%) of 25-way multiple-choice activity recognition on 500 four-frame clips on BARISTA, an egocentric coffee-preparation video dataset with verified per-frame scene graphs, zero-shot with shared task prompts, temperature 0.7 (GPT-5.4 mini at its fixed sampling), low thinking effort for the closed models, mean of three runs; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 27B | 77.8 |
| 2 | Qwen 3.5 9B | 71.6 |
| 3 | Gemma 4 31B | 68.4 |
| 4 | GPT-5.4 Mini (Low) | 63.1 |
Interactive version: theaggregate.ai/benchmark?slug=barista-activity-recognition · How It Works · Data refreshed daily, snapshot 2026-10-07.