ai.paski.devLlama 4 Maverick Instruct
llmA natively multimodal mixture-of-experts model with 17B active parameters routed across 128 experts (400B total), Meta's flagship Llama 4 release.
Metadata
| org | Meta |
| released | 2025-04-05 |
| params | 400B-A17B |
| license | llama-4 |
| kind | llm |
| hf id | meta-llama/Llama-4-Maverick-17B-128E-Instruct |
Benchmarks
| MMLU-Pro | 80.5 | source ↗ |
| GPQA | 69.8 | source ↗ |
Lineage
No recorded lineage — root or standalone model.
view in full graph →