Title: Llama 4 Scout 17B-16E Instruct vs Llama 4 Maverick 17B %E2%80%94 LLM EU
Canonical: https://llmeu.com/cs/compare/llama-4-scout-17b-16e-instruct-vs-llama-4-maverick-17b-128e-instruct/
Markdown mirror: https://llmeu.com/cs/compare/llama-4-scout-17b-16e-instruct-vs-llama-4-maverick-17b-128e-instruct.md
Machine-readable: https://llmeu.com/llms.txt (index) - https://llmeu.com/llms-full.txt (full catalogue) - https://llmeu.com/llms-catalogue.json

# Llama 4 Scout 17B-16E Instruct vs Llama 4 Maverick 17B-128E Instruct

Porovnejte dva a%C5%BE %C4%8Dty%C5%99i modely vedle sebe podle pol%C3%AD, kter%C3%A1 rozhoduj%C3%AD o nasazen%C3%AD: kontext, cena, region, licence, jazyky, hosting a co je ov%C4%9B%C5%99eno.

Tam, kde se dva modely li%C5%A1%C3%AD, tabulka uk%C3%A1%C5%BEe rozd%C3%ADl a skon%C4%8D%C3%AD. LLM EU nezve%C5%99ej%C5%88uje odznak v%C3%ADt%C4%9Bze bez dolo%C5%BEen%C3%A9ho sk%C3%B3re pro danou %C3%BAlohu a chyb%C4%9Bj%C3%ADc%C3%AD %C3%BAdaj z%C5%AFst%C3%A1v%C3%A1 chyb%C4%9Bj%C3%ADc%C3%AD, m%C3%ADsto aby se stal nulou.

## Kdy pou%C5%BE%C3%ADt

### [Llama 4 Scout 17B-16E Instruct](/cs/models/llama-4-scout-17b-16e-instruct/)

#### Siln%C3%A9 str%C3%A1nky

- 10M-token context window, the longest published in the Llama line
- multimodal input with image grounding
- fits a single H100 at Int4 quantisation

#### Omezen%C3%AD

- Llama 4 Community License, not an OSI licence
- 109B total parameters make full-precision serving expensive
- long-context quality at the 10M limit is not independently verified

#### %C3%9Alohy

[Zpracov%C3%A1n%C3%AD obrazu](/cs/tasks/vision/), [Shrnut%C3%AD](/cs/tasks/summarization/), [Vyhled%C3%A1v%C3%A1n%C3%AD (RAG)](/cs/tasks/rag/), [V%C3%ADcejazy%C4%8Dn%C3%A1 podpora](/cs/tasks/multilingual-support/)

### [Llama 4 Maverick 17B-128E Instruct](/cs/models/llama-4-maverick-17b-128e-instruct/)

#### Siln%C3%A9 str%C3%A1nky

- multimodal input with 1M-token context
- 17B active parameters keep per-token compute moderate
- vendor reports competitiveness with much larger dense models

#### Omezen%C3%AD

- about 400B total parameters must fit in memory even though only 17B are active
- Llama 4 Community License, not an OSI licence
- Meta%E2%80%99s benchmark comparisons are self-reported

#### %C3%9Alohy

[Zpracov%C3%A1n%C3%AD obrazu](/cs/tasks/vision/), [Uva%C5%BEov%C3%A1n%C3%AD](/cs/tasks/reasoning/), [V%C3%ADcejazy%C4%8Dn%C3%A1 podpora](/cs/tasks/multilingual-support/), [Shrnut%C3%AD](/cs/tasks/summarization/)

**%C3%9Alohy:** [Zpracov%C3%A1n%C3%AD obrazu](/cs/tasks/vision/), [Shrnut%C3%AD](/cs/tasks/summarization/), [V%C3%ADcejazy%C4%8Dn%C3%A1 podpora](/cs/tasks/multilingual-support/)

## Porovnat modely

- [Bloom vs Llama 4 Scout 17B-16E Instruct](/cs/compare/bloom-vs-llama-4-scout-17b-16e-instruct/)
- [Command A+ (05-2026) vs Llama 4 Maverick 17B-128E Instruct](/cs/compare/command-a-plus-05-2026-vs-llama-4-maverick-17b-128e-instruct/)
- [DeepSeek R1 0528 vs Llama 4 Maverick 17B-128E Instruct](/cs/compare/deepseek-r1-0528-vs-llama-4-maverick-17b-128e-instruct/)
- [DeepSeek-R1 vs Llama 4 Maverick 17B-128E Instruct](/cs/compare/deepseek-r1-vs-llama-4-maverick-17b-128e-instruct/)
- [DeepSeek V3 0324 vs Llama 4 Maverick 17B-128E Instruct](/cs/compare/deepseek-v3-0324-vs-llama-4-maverick-17b-128e-instruct/)
- [DeepSeek V3.1 vs Llama 4 Maverick 17B-128E Instruct](/cs/compare/deepseek-v3-1-vs-llama-4-maverick-17b-128e-instruct/)