Meta Llama API pricing: all models, per million tokens
Input and output prices per million tokens, context windows and retirement dates for 8 Meta Llama models, from the OpenRouter catalog. Updated automatically.
| Model | Input / 1M | Output / 1M | Context |
|---|---|---|---|
| meta-llama/llama-guard-4-12b | US$ 0.18 | US$ 0.18 | 163,840 |
| meta-llama/llama-4-maverick | US$ 0.35 | US$ 1.00 | 524,288 |
| meta-llama/llama-4-scout | US$ 0.1 | US$ 0.3 | 1,310,720 |
| meta-llama/llama-3.3-70b-instruct | US$ 0.71 | US$ 0.71 | 131,072 |
| meta-llama/llama-3.2-1b-instruct retiring 2026-11-20 | US$ 0.027 | US$ 0.201 | 60,000 |
| meta-llama/llama-3.2-3b-instruct | US$ 0.05 | US$ 0.33 | 131,072 |
| meta-llama/llama-3.1-70b-instruct | US$ 0.72 | US$ 0.72 | 131,072 |
| meta-llama/llama-3.1-8b-instruct | US$ 0.02 | US$ 0.04 | 131,072 |
Reference prices from OpenRouter (aggregator); they may differ from the provider's direct pricing.
Give your AI agent these facts
Vigia is a free MCP server: your coding agent checks versions, breaking changes and vulnerabilities before it writes code. No signup, no API key.
claude mcp add --transport http vigia https://vigia.coredls.cloud/mcp
Add to CursorAdd to VS Code★ Star on GitHub
Any other MCP client
{ "mcpServers": { "vigia": { "url": "https://vigia.coredls.cloud/mcp" } } }Agent skill (SKILL.md): tells the agent when to use Vigia · GitHub Action for pull requests
For agents
GET https://vigia.coredls.cloud/v1/models?provider=meta-llama