Meta
Llama 4 Scout
Open SourceMeta's ultra-long-context open-weights model with a 10M token window — the largest of any publicly available model. Scout is a smaller MoE variant (109B total, ~17B active) optimized for speed and context length over raw intelligence. At 135 t/s and AA Intelligence Index 14, it's the right call when you need to process enormous documents or codebases that would overflow any other model.
Context window
10.0M tokens
API (blended)
$0.17/1M
Consumer access
Free (limited)
Multimodal
Yes
Score Breakdown
39.1/100 → 3.9/10Intelligence, Reliability, Speed, and Context are field-relative — scores shift as models are added. Accessibility and Trust are absolute checklists. Full methodology →
Strengths
- +10M token context — largest of any publicly available model; processes entire repositories or books in one call
- +Open weights (Llama 4 Community License) — fully self-hostable and auditable
- +135 t/s (AA-measured) — fastest of any model reviewed, including much smaller ones
- +Ultra-cheap API: ~$0.17/1M blended via Groq — cheapest in the dataset
- +Natively multimodal: image + text input
- +Active ecosystem: Groq, Together, Fireworks, Ollama, and LM Studio support
Weaknesses
- -AA Intelligence Index 13 — the lowest in this dataset; not suitable for complex reasoning tasks
- -GPQA Diamond 58.7%, HLE 4.3% (AA-measured) — near the lower bound for frontier science tasks
- -τ²-bench 15.5% (AA-measured) — very limited agentic tool use capability
- -Meta data practices: not ideal for privacy-sensitive enterprise use
- -10M context not available through all hosted providers — check limits before integrating
- -No MCP support
Best for
Not ideal for
Pricing details
Subscription plans
API pricing
Prices verified February 2026. LLM pricing changes frequently — verify at the provider's site before budgeting.
Last updated: February 27, 2026