Model comparison
Llama 4 Scout vs Claude Sonnet 4.6
Meta
Llama 4 Scout
Open-weights frontier with a headline 10M-token context.
Anthropic
Claude Sonnet 4.6
The pragmatic default — Claude quality without Opus pricing.
Specs
| Metric | Llama 4 Scout | Claude Sonnet 4.6 |
|---|---|---|
| Context window | 328K tokens | 1M tokens↑ |
| Input $/1M tokens | $0.080↑ | $3.00 |
| Output $/1M tokens | $0.300↑ | $15.00 |
| Modalities | Text · Image | Text · Image |
| Open weights | Yes | No |
| Released | Apr 2025 | — |
Capability differences
| Capability | Llama 4 Scout | Claude Sonnet 4.6 |
|---|---|---|
| Extended thinking | No | Yes |
| Prompt caching | No | Yes |
| Open weights | Yes | No |
Key differences
- —Llama 4 Scout offers 328K context; Claude Sonnet 4.6 offers 1M.
- —Input pricing: Llama 4 Scout at $0.080/1M vs Claude Sonnet 4.6 at $3.00/1M.
- —Llama 4 Scout is open weights; Claude Sonnet 4.6 is proprietary.