LLM WORLD

Llama 4 Scout

Meta AI · · 192 days after Llama 3.2

Verified

3 of 3 published values traced to a source2 primary sources, 2 archived

Why this status

Split from the former "llama-4" record, which was named for the family while holding one variant's figures. Evidence: "Llama 4 Scout, is a general purpose model with 17 billion active parameters, 16 experts, and 109 billion total parameters... Llama 4 Scout dramatically increases the supported context length from 128K in Llama 3 to an industry leading 10 million tokens." Capabilities evidenced from the primary sources: vision: “As our product workhorse model for general assistant and chat use cases, Llama 4 Maverick is great for precise image understanding and creative writing.”.

What these states mean · How records are judged · Has this record changed?

Meta's smaller Llama 4 mixture-of-experts model, with a 10-million-token context window.

visionopen-weightsmultimodallong-context

Details

ReleasedApril 5, 2025
stated in ai.meta.com, huggingface.co

Llama 4 Scout was released on April 5, 2025.

CompanyMeta AI
FamilyLlama
Released alongsideLlama 4 Maverick
EraReasoning models
TypeLanguage model · LLM
ModalitiesText + Image → Text
CapabilitiesVision
WeightsOpen weights
LicenceLlama 4 Community License
Context window10M tokens
stated in ai.meta.com, huggingface.co

Llama 4 Scout supports a context window of 10,000,000 tokens.

Parameters109B
stated in ai.meta.com, huggingface.co

Llama 4 Scout has 109,000,000,000 parameters.

Current information, elsewhere

This record is a historical one: every figure above is traced to a dated primary source. For what is true today — live pricing, current rankings, availability — these are better maintained than anything here could be.

Sources

Record status: verified · confidence 90/100 · 2 of 2 primary

Split from the former "llama-4" record, which was named for the family while holding one variant's figures. Evidence: "Llama 4 Scout, is a general purpose model with 17 billion active parameters, 16 experts, and 109 billion total parameters... Llama 4 Scout dramatically increases the supported context length from 128K in Llama 3 to an industry leading 10 million tokens." Capabilities evidenced from the primary sources: vision: “As our product workhorse model for general assistant and chat use cases, Llama 4 Maverick is great for precise image understanding and creative writing.”.

Primary means published by the organisation that made the model. Verified records require at least one.

Llama family

  1. Llama 1 February 24, 2023
  2. Llama 2 July 18, 2023
  3. Llama 3 April 18, 2024
  4. Llama 3.1 July 23, 2024
  5. Llama 3.2 September 25, 2024
  6. Llama 4 Scout April 5, 2025
  7. Llama 4 Maverick April 5, 2025
  8. Llama 5 April 8, 2026

Browse models

Compare Llama 4 Scout with another model →
See Llama 4 Scout on the interactive timeline →