BARISTA - Referring Expression Generation: leaderboard

BARISTA (2026) is an egocentric benchmark of 185 coffee-preparation videos with verified per-frame scene graphs, from which zero-shot language tasks for vision-language models are derived. This board is the correctness of a generated description of a marked region, judged by Gemini 2.5 Flash on a 0 to 1 scale; higher is better.

Metric: Correctness of the generated description of a marked region, Gemini 2.5 Flash G-Eval judge (0 to 10 rubric rescaled to 0-1), over 1,000 frozen examples 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 February 2027. 5 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 FlashLow0.76#175
2Gemma 4 31B0.667#382
3Qwen 3.5 27B0.622#375
4Qwen 3.5 9B0.576#828
5GPT-5.4 MiniLow0.308#406

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=barista-referring-expression-generation · How It Works · Data refreshed daily, snapshot 2026-11-21.