AVoCaDO: benchmark results
Provider: Other.
Aggregate Rating 1671 ± 43, rank #496 of 10916 rated models, from 8 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGC-VideoCap - Average | 73.2 | Mean judge score (audio, visual, detail) | 100 |
| CapRiCorn-1K | 28.8 | Keypoint accuracy (%; share of annotated keypoints a GPT-4.1 | 88.2 |
| CapRiCorn-1K - Referential Consistency | 18.4 | Subject referential consistency (%; share of same-subject de | 88.2 |
| video-SALMONN-2 Test Set (GPT-4.1 judge) - Total error | 37.3 | Missed + hallucinated events (%) | 83.3 |
| Daily-Omni (caption QA, Gemini-2.5-Pro reader) | 50.1 | Accuracy (%) | 80 |
| WorldSense (caption QA, Gemini-2.5-Pro reader) | 25.7 | Accuracy (%) | 80 |
| DiaDemBench - Speaker attribution F1 (REF) | 38.7 | REF F1 | 61.5 |
| DiaDemBench - Utterance transcription F1 (ASR) | 51.7 | ASR F1 | 61.5 |
Interactive version: theaggregate.ai/model?slug=avocado · How It Works · Data refreshed daily, snapshot 2026-11-08.