Llama 4 Scout
Released 2025-04 · chat · 10.0M tokens · 6 benchmarks · Open weight
Editorial notes
MoE 17B activos, 16 expertos, 109B totales. Context window record de 10M tokens.
Spec sheet
- Company
- Meta
- Country
- US
- Type
- chat
- Release
- 2025-04
- Context
- 10.0M tokens
- Params total
- 109B
- Params active (MoE)
- 17B
- License
- llama-4-community
- Quants
- BF16, Q8_0, Q5_K_M, Q4_K_M
- Slug
- llama-4-scout
Quick install
3 toolsollama run llama4:scoutBuscar "Llama-4-Scout" en LM Studiovllm serve meta-llama/Llama-4-Scout-17B-16E-Instruct --max-model-len 1000000Note: Q4_K_M ~62 GB VRAM. Context window de 10M tokens excepcional.
Benchmarks (6)
Reasoning 4
- 79.6MMLUMassive Multitask Language Understanding — 57 academic subjects, ~16K questions.
- 74.3MMLU-ProMMLU upgraded with harder questions and 10 answer options.
- 69.4MMMUMultimodal Multidiscipline Understanding — multimodal reasoning over academic im
- 57.2GPQA-DiamondGraduate-level Physics, Chemistry, Biology — PhD-level questions.
Multilingual 1
Cite this model
BibTeX · APA
BibTeX
@misc{frontier-llama-4-scout,
title = {Llama 4 Scout},
author = {{Meta}},
year = {2025},
note = {Frontier Benchmarks AI atlas. Accessed 2026-05-08},
url = {https://frontierbenchmarks.com/models/llama-4-scout}
} APA
Meta (2025). Llama 4 Scout [Large language model]. Frontier Benchmarks AI. Retrieved 2026-05-08, from https://frontierbenchmarks.com/models/llama-4-scout
Citation reflects the atlas page, not the original model paper. For the paper, see the "Resources" section above.