Llama 4 Scout

Overview

Llama 4 Scout is a natively multimodal model capable of processing both text and images. It features a 17 billion activated parameter (109B total) mixture-of-experts (MoE) architecture with 16 experts, supporting a wide range of multimodal tasks such as conversational interaction, image analysis, and code generation. The model includes a 10 million token context window.

Llama 4 Scout was released on April 5, 2025. API access is available through 6 providers, including DeepInfra, Lambda and others.

Benchmarks

Llama 4 Scout Performance Across Datasets

Scores sourced from the model's scorecard, paper, or official blog posts

llm-stats.com - Wed Dec 10 2025

Pricing

Pricing, performance, and capabilities for Llama 4 Scout across different providers:

Provider	Input ($/M)	Output ($/M)	Max Input	Max Output	Latency (s)	Throughput	Quantization	Input	Output
DeepInfra	$0.08	$0.30	10.0M	10.0M	0.31	76.1 tok/s	—	Text Image Audio Video	Text Image Audio Video
Lambda	$0.08	$0.30	10.0M	10.0M	0.43	139.7 tok/s	—	Text Image Audio Video	Text Image Audio Video
Novita	$0.10	$0.50	10.0M	10.0M	0.85	69.82 tok/s	—	Text Image Audio Video	Text Image Audio Video
Groq	$0.11	$0.34	10.0M	10.0M	1.08	776.1 tok/s	—	Text Image Audio Video	Text Image Audio Video
Fireworks	$0.15	$0.60	10.0M	10.0M	0.53	116.1 tok/s	—	Text Image Audio Video	Text Image Audio Video
Together	$0.18	$0.59	10.0M	10.0M	0.54	106.9 tok/s	—	Text Image Audio Video	Text Image Audio Video

Price Comparison for Llama 4 Scout

Price per 1M input tokens (USD), lower is better

llm-stats.com - Wed Dec 10 2025

Throughput Comparison for Llama 4 Scout

Tokens per second, higher is better

llm-stats.com - Wed Dec 10 2025

Latency Comparison for Llama 4 Scout

Time to first token (s), lower is better

llm-stats.com - Wed Dec 10 2025

Example Outputs

API Access