Model comparison
Llama 3.2 3B Instruct (free) vs Llama 4 Scout
Compare Llama 3.2 3B Instruct (free) and Llama 4 Scout: input/output $/Mtoken, context window, modalities, and license. OpenRouter-synced pricing.
Meta
Llama 3.2 3B Instruct (free)
Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summariza
Meta
Llama 4 Scout
Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B. It supports native multimo
| Metric | Llama 3.2 3B Instruct (free) | Llama 4 Scout |
|---|---|---|
| Provider | Meta | Meta |
| Context window | 131,072 | 10,000,000 |
| Input $/Mtok | $0 | $0.100 |
| Output $/Mtok | $0 | $0.300 |
| Modalities | text | text, vision |
| License | open-weights | open-weights |
Quick take
On input price, Llama 3.2 3B Instruct (free) is cheaper at $0/Mtok. For context window, Llama 4 Scout leads with 10,000,000 tokens.
Pick based on your workload: high-volume cheap inference vs long-document / agent loops. Enterprise DNA can wire either model into Omni with evals, secrets, and job orchestration.