aya-vision-8B — benchmark results

Cohere Aya Vision 8B multimodal model for multilingual vision-language tasks. Provider: Cohere. Released 2025-03-02. Access: Open.

Unified ELO 1368 ± 21, rank #1352 of 1776 rated models, from 49 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LatamBoard - Spanish TeleIA71.43Score (%)86.4
LatamBoard - Spanish MGSM12Score (%)72.7
LatamBoard - FaQuAD NLI74.19Score (%)69.7
EVALITA - word-in-context63.86CPS61.7
JMMMU-Pro - Japanese Heritage30.67Accuracy (%)61.5
LatamBoard - ASSIN2 RTE91.06Score (%)60.6
LatamBoard - Portuguese Score85.93Score (%)60.6
KOFFVQA - Recognition75Score (%)58
KOFFVQA - Korean Recognition25Score (%)51.9
LatamBoard - ENEM Challenge65.85Score (%)51.5
LatamBoard - OAB Exams46.61Score (%)51.5
LatamBoard - Spanish Score59.79Score (%)51.5

Interactive version: theaggregate.ai/model?slug=aya-vision-8b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.