Model Terminal

Llama 4 Scout

Llama 4 Scout is an open-weight, instruction-tuned multimodal language model released by Meta on April 5, 2025, featuring 17 billion active parameters across 16 experts (mixture-of-experts architecture) and a 10-million-token context window. It accepts text and image inputs and is designed for assistant-style tasks including question answering, summarization, image captioning, and long-context reasoning. Source: supabase.

Value score
Not scored
Context
10M
tokens
Max output
tokens
Price

Capability radar

Identity

Developer
meta-llama
Openness
Open Weights
Modalities
Release
2025-04-05
Knowledge cutoff
Deprecation

Benchmarks

No benchmark scores yet.

Compare nearby

Pricing

Input / 1M
Output / 1M
Speed