Llama 4 Maverick token cost & pricing
Llama 4 Maverick is Meta's open-weight mixture-of-experts model with a 1M context window. Because the weights are open, pricing depends entirely on the host — the figures here reflect typical third-party serverless rates rather than a single vendor's list price.
Llama 4 Maverick specifications
| Vendor | Meta |
|---|---|
| Input price | $0.27 / 1M tokens |
| Output price | $0.85 / 1M tokens |
| Context window | 1,048,576 tokens |
| Max output | 16,384 tokens |
| Open weights | Yes |
| Price rank | 4 of 17 (cheapest first) |
| Best for | open-weights, long-context |
What Llama 4 Maverick costs in practice
Four common workloads, priced at list rates. Monthly figures assume 30 days.
| Workload | Per request | Requests/day | Per month |
|---|---|---|---|
| Chatbot · 1K in / 500 out | $0.0007 | 10,000 | $208.50 |
| RAG Q&A · 8K in / 400 out | $0.0025 | 5,000 | $375.00 |
| Doc summary · 30K in / 1K out | $0.0089 | 1,000 | $268.50 |
| Code agent · 60K in / 3K out | $0.02 | 500 | $281.25 |
How many requests does $100 buy?
Assuming a 1,000-token prompt and a 500-token response per request.
How Llama 4 Maverick pricing actually works
At roughly $0.27 per million input and $0.85 per million output from mainstream hosts, Maverick is among the cheapest ways to get a 1M-token window. Costs vary widely between providers, and self-hosting on your own GPUs can beat these rates at sustained high volume.
When to use Llama 4 Maverick
Good for long-document processing, bulk summarization and multilingual work where you want to avoid vendor lock-in or need on-premise deployment for compliance. Quality trails the closed frontier models on complex reasoning.
How to cut your Llama 4 Maverick bill
Compare hosts before committing — the spread on open-weight serving is large. If your volume is steady and high, reserved capacity or self-hosting usually undercuts per-token serverless pricing substantially.
Cheaper alternatives to Llama 4 Maverick
Run the numbers on your own prompt
Paste your actual text into the token counter to get an exact count instead of an estimate. Nothing is uploaded.
Open the calculator →