Llama 4 Scout

Flagship

by Meta

Llama 4 Scout is Meta's open-source model with the largest context window of any production model at 10 million tokens. Free to download and self-host, it opens up use cases that were previously impossible due to context limitations.

Input Price

Free

per 1M tokens

Output Price

Free

per 1M tokens

Context Window

10M

tokens

Released

2025-04

Open source

Capabilities

textvisioncode

Key Strengths

  • 10M token context window
  • Fully open source
  • Free to self-host
  • Vision capable

Best For

  • Massive document processing
  • Entire repository analysis
  • Self-hosted deployments
  • Research and experimentation

Benchmark Scores

BenchmarkScoreDescription
MMLU-Pro85.9General knowledge and reasoning across 57 subjects
HumanEval88.4Python code generation and problem solving
GPQA Diamond56.2Graduate-level science questions verified by domain experts
MATH79.8Competition-level mathematics problems
SWE-bench44.6Real-world software engineering tasks from GitHub issues

Scores sourced from public benchmark datasets. See full benchmark leaderboard for all models.

Open Source Model

Llama 4 Scout is free to download and self-host under the Llama 4 Community License. Hosted API pricing varies by provider (e.g., Together, Fireworks, Groq). See our open source LLM guide for deployment options.

Related Models

View DocumentationCompare ModelsCost CalculatorFull Pricing Guide