Mistral Large 4 vs Mistral Large 3: pricing compared
For a 30-step coding agent run, Mistral Large 3 is 1.4x cheaper ($0.21 vs $0.30). Prices verified 2026-11-17.
| Mistral Large 4 | Mistral Large 3 | |
|---|---|---|
| Provider | Mistral AI | Mistral AI |
| Input / 1M | $0.70 | $0.50 |
| Cached input / 1M | $0.07 | $0.05 |
| Output / 1M | $2.10 | $1.50 |
| Batch discount | 50% | 50% |
| Context window | 1M | 256K |
| Long-context tier | none | none |
Cost on real workloads
| Workload | Mistral Large 4 | Mistral Large 3 |
|---|---|---|
| 10,000 chatbot replies 2,000 input + 400 output tokens each, 50% of the prompt cached | $16.10 | $11.50 |
| One 30-step coding agent run 12K system + tools prompt, 1K output and 3K tool result per step, caching on | $0.30 | $0.21 |
| 1M documents classified (batch) 600 input + 20 output tokens each, batch API where available | $231 | $165 |
Price is only half the decision: a cheaper model that needs more steps or retries can cost more per finished task. Run your own numbers in the calculator (both models pinned), or see how both rank in the use-case guides. Every other model is in the pricing comparison, and the token counter shows how many tokens your prompt is on each.
FAQ
Is Mistral Large 4 cheaper than Mistral Large 3?
For a typical 30-step coding agent run with caching, Mistral Large 3 costs $0.21 and Mistral Large 4 costs $0.30, so Mistral Large 3 is 1.4x cheaper. Per 1M tokens, Mistral Large 4 is $0.70 / $2.10 and Mistral Large 3 is $0.50 / $1.50.
Which has the bigger context window?
Mistral Large 4: 1,000,000 tokens. Mistral Large 3: 256,000 tokens.