Llama 3.2 90B — benchmark results

Meta's 90B vision-language base model from Llama 3.2, adding cross-attention image adapters to Llama text weights with a 128K context. Provider: Meta. Released 2024-09-25. Access: Open.

Unified ELO 1457 ± 27, rank #965 of 1776 rated models, from 21 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Translation Set1 to en spBleu43.7Set1→en spBLEU (self-reported)75
AI for Education Pedagogy - Technology82.08Accuracy (%)71.6
MMMU Benchmark60.3Validation Score66.7
AI for Education Pedagogy - Maths76.19Accuracy (%)50.9
AI for Education Pedagogy76.31Accuracy (%)49.1
AI for Education Pedagogy - Secondary75Accuracy (%)47.5
AI for Education Pedagogy - Social studies74.55Accuracy (%)47
AI for Education SEND70.64Accuracy (%)44.6
AI for Education Pedagogy - Primary79.81Accuracy (%)42.7
AI for Education Pedagogy - Science74.86Accuracy (%)38.6
AI for Education Visual Maths - Geometry36.92Accuracy (%)34.2
AI for Education Visual Maths - Statistics and Probability14.29Accuracy (%)29.2

Interactive version: theaggregate.ai/model?slug=llama-3-2-90b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.