IDEFICS-80B-Instruct: benchmark results
Provider: HuggingFace. Released 2023-08-22. Access: Open.
Aggregate Rating 1251 ± 30, rank #9248 of 10916 rated models, from 364 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenVLM COCO Captions | 94.9 | CIDEr | 98.1 |
| HELM VHELM - Blink Perception - Art Style | 100 | Prefix Quasi-Exact Match (%) | 96.8 |
| HELM VHELM - Blink Reasoning - Jigsaw | 98 | Prefix Quasi-Exact Match (%) | 86.2 |
| OpenVLM MME - Scene | 163.2 | Score | 82.9 |
| OpenVLM MMMU - Music | 36.7 | Accuracy (%) | 75.4 |
| HELM VHELM - Blink Reasoning - Multi-view Reasoning | 72.93 | Prefix Quasi-Exact Match (%) | 57.4 |
| HELM VHELM - Safety | 55.32 | Mean win rate | 55.3 |
| HELM VHELM - Image2Webpage | 0.26 | EMD-Sim | 52.9 |
| OpenVLM MathVista - LOG | 18.9 | Accuracy (%) | 49.3 |
| OpenVLM MM-Vet - Knowledge | 38.6 | Score | 48.2 |
| OpenVLM MM-Vet - Language Generation | 37.1 | Score | 47.5 |
| OpenVLM MME - Posters | 158.5 | Score | 46.2 |
Interactive version: theaggregate.ai/model?slug=idefics-80b-instruct · How It Works · Data refreshed daily, snapshot 2026-11-08.