Llama 4 Scout — reviews, specs & pricing
Efficient Llama 4 MoE: 109B total / 17B active parameters with a 10M-token context window.
Summary
Llama 4 Scout by Meta is an efficient Mixture of Experts (MoE) chat model featuring 109 billion total and 17 billion active parameters. It stands out in the landscape with an industry-leading 10-million-token context window, allowing it to process massive documents and long-form conversations seamlessly. This model successfully bridges the gap between high-performance reasoning and cost-effective, chat-optimized deployment.
Sample use case
An enterprise legal department can deploy Llama 4 Scout to review and analyze entire case histories, compliance archives, and multi-volume contracts simultaneously. By loading these extensive documents directly into the model's massive 10-million-token context window, lawyers can query the chat interface to instantly identify discrepancies, summarize precedent cases, and draft precise briefs without needing to fragment the data or rely on external retrieval systems.
Specifications
- Provider: meta
- License: source-available
- Parameters: 109B
- Context: 10000k tokens
- Released: 2025-04-05
Pros
- Massive 10-million-token context window
- Highly efficient MoE architecture with 17B active parameters
- Cost-effective deployment for enterprise chat applications
- Robust long-context reasoning capabilities
Cons
- Requires significant hardware to host the 109B total parameter size
- Optimized only for text and chat modalities
- Potential inference latency spikes when utilizing the full context limit
Average rating 0.0 from 0 community reviews on Reviuws.