Llama 4 Maverick Instruct: benchmark results
Meta's natively multimodal open MoE (400B total/17B active, 128 experts) with a 1M-token context (April 2025). Provider: Meta. Released 2025-04-05. Access: Open.
Unified ELO 1564 ± 1, rank #386 of 1392 rated models, from 125 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SEA LLM Leaderboard - SeaExam | 76.7 | Private Average Score (%) | 92.2 |
| KOFFVQA - Relationship | 74.33 | Score (%) | 85.2 |
| KOFFVQA - Graph and Chart Understanding | 83.33 | Score (%) | 84.6 |
| KOFFVQA - Hallucination and Robustness | 80 | Score (%) | 80.9 |
| KOFFVQA - Object Attributes | 79.17 | Score (%) | 79 |
| KOFFVQA - Document Understanding | 86 | Score (%) | 78.4 |
| MERA v2 - Enantiosemy | 49.6 | Score (%) | 77.8 |
| Vals AI MGSM | 92.44 | Accuracy (%) | 77.3 |
| Enkrypt AI - Jailbreak Risk | 4.54 | Risk Score | 76.9 |
| UGI - Natural Intelligence | 30.93 | NatInt Score | 76 |
| ProLLM - LLM-as-a-Judge | 80.9 | Score (%) | 75.9 |
| KOFFVQA | 76.95 | Overall Score (%) | 75.3 |
Interactive version: theaggregate.ai/model?slug=llama-4-maverick-instruct · How It Works · Data refreshed daily, snapshot 2026-09-05.