For coding agents, Llama 4 Maverick and Codestral cost about the same per request ($0.0072 vs $0.0070 on our measured traffic).
Run current Claude, GPT, Gemini and open models in your agent
Or let the router send each step to the cheapest model that can do it, on providers we vet for price, uptime and speed. Plans from $19 a month.
| Cost at list price | Codestral | Llama 4 Maverick |
|---|---|---|
| Prompt under 100K tokens50% of requests, 87% of input cached | $0.0043 | $0.0043 |
| Prompt of 100K to 272K39% of requests, 88% of input cached | $0.011 | $0.011 |
| Prompt over 272K tokens11% of requests, 86% of input cached | n/a | n/a |
| Average request | $0.0070 | $0.0072 |
| 100 requests | $0.70 | $0.72 |
| Median session16 requests; a session ends after 10 idle minutes | $0.084 | $0.080 |
| A month of regular usemedian of 20 users active on 5+ days | $8.91 | $9.25 |
Each logged request priced at the listed rates above for its own token counts, including cache reads and any higher rate for long prompts. Average output: 698 tokens per request. Both models are priced on the requests up to 256K tokens that fit in both context windows; band shares above describe all requests.
Codestral's 256K window cannot take 13% of coding-agent requests, so the comparison uses only requests both can take. 87% of agent input is read from cache, and Llama 4 Maverick charges $0.05 per million for it against $0.03. The price is OpenRouter's listed rate on October 13, 2026, which can be a third-party host's; Meta's own API may charge differently.
Pick a winner in each dimension. You can change your vote any time.
Correctness and depth of what it produces
Tool calls, instruction-following, and multi-step tasks
Tokens per second and time-to-first-token
How much capability you get per dollar
Consistent results: fewer refusals, loops and format breaks
The agent matters as much as the model. Compare AI agents →
Current Claude, GPT, Gemini and open models run on Standard Compute