Skip to main content
MetaReleased April 5, 2025

Llama 4 Scout

Open Source#12 Arena Rank109B (17B active) parameters

Llama 4 Scout holds a solid spot in the Arena rankings at #12. Context window: 0.01K tokens.

Context

10M

Input

Free

Key Specifications

🏆

Arena Rank

#12

📐

Context Window

10M

📥

Input Price

per 1M tokens

Free

📤

Output Price

per 1M tokens

Free

🧠

Parameters

109B (17B active)

🔓

Open Source

Yes

Best For

Long contextopen sourcemultilingual

About Llama 4 Scout

Llama 4 Scout, developed by Meta AI, is a Mixture-of-Experts model designed for efficient deployment with strong performance across general reasoning, coding, and multilingual tasks. The model uses sparse expert routing to maintain high capability while reducing inference compute requirements. As part of Meta's Llama 4 family, Scout represents the efficiency-optimized variant, targeting developers who need capable AI at manageable computational costs. The model supports long context processing and demonstrates improved instruction following compared to Llama 3 series models. Free and open-source under Meta's license, it can be deployed on enterprise hardware without API costs. Llama 4 Scout continues Meta's commitment to open-source AI development, providing the community with a model that balances capability and deployment practicality for production applications at scale.

Pricing per 1M tokens

Input Tokens

Free

Output Tokens

Free

Compare Llama 4 Scout

See how Llama 4 Scout stacks up against other leading AI models

Llama 4 Scout vs Llama 4 MaverickCompare →Llama 4 Scout vs Llama 3.3Compare →Llama 4 Scout vs Claude Opus 4Compare →Llama 4 Scout vs GPT-o3Compare →Llama 4 Scout vs GPT-4oCompare →

Other Meta Models

Llama 4 Maverick
1M#7
Llama 3.3
128K#13

Other Top Models

Claude Opus 4Anthropic
#1
GPT-o3OpenAI
#2
GPT-4oOpenAI
#2
GPT-o1OpenAI
#3
DeepSeek R1DeepSeek
#3
Claude Sonnet 4Anthropic
#3

Explore More

Model Benchmarks →All Models →

Frequently Asked Questions

What is Llama 4 Scout?▾
Llama 4 Scout, developed by Meta AI, is a Mixture-of-Experts model designed for efficient deployment with strong performance across general reasoning, coding, and multilingual tasks. The model uses sparse expert routing to maintain high capability while reducing inference compute requirements. As part of Meta's Llama 4 family, Scout represents the efficiency-optimized variant, targeting developers who need capable AI at manageable computational costs. The model supports long context processing and demonstrates improved instruction following compared to Llama 3 series models. Free and open-source under Meta's license, it can be deployed on enterprise hardware without API costs. Llama 4 Scout continues Meta's commitment to open-source AI development, providing the community with a model that balances capability and deployment practicality for production applications at scale.
How much does Llama 4 Scout cost?▾
Input pricing for Llama 4 Scout is Free per million tokens; output runs Free. No cost at all.
What is Llama 4 Scout's context window?▾
The context window for Llama 4 Scout is 10M tokens. That's the maximum amount of text you can feed into a single prompt, including system instructions, conversation history, and the actual query.
Is Llama 4 Scout open source?▾
Llama 4 Scout is fully open source. You can grab the weights, run it on your own hardware, and fine-tune it for specific tasks. That flexibility is a big deal for teams with strict data requirements.
What is Llama 4 Scout best for?▾
The sweet spot for Llama 4 Scout is: Long context, open source, multilingual. If your workload fits one of these categories, it's worth benchmarking against alternatives.