Codestral 22B by Mistral AI costs $0.300 per 1 Million input tokens and $0.900 per 1 Million output tokens. It features a 128k context window and has a blended 3:1 production rate of $0.450/1M tokens.
Codestral 22B
Open-weight coding specialist model fluent in 80+ programming languages.
Verified Specifications & Benchmark Data
| Developer / Provider | Mistral AI |
| Model Family | Mistral |
| Blended 3:1 Rate (Production Benchmark) | $0.450 / 1M tokens |
| Batch API Discount (24hr SLA) | None |
| Multimodal Vision | ❌ Text Only |
| Function Calling / Structured Outputs | ✅ Native Tool Calling |
| MMLU Benchmark Score | 81.1% |
| Latency & Throughput Tier | Ultra-Fast (~80 t/s) |
| Optimal Architecture & Use Cases | Fill-in-the-middle code completion, automated unit tests, IDE integration |
Frequently Asked Questions about Codestral 22B
How much does Codestral 22B cost per 1M tokens?
Codestral 22B pricing is set at $0.300 per 1 million input tokens and $0.900 per 1 million output tokens. For high-volume batch workloads, 24-hour batch queues provide a None.
How much money does Codestral 22B prompt caching save?
Prompt caching is currently not natively offered for Codestral 22B on direct serverless endpoints.
What is the context window limit of Codestral 22B?
Codestral 22B has a maximum context window of 128k (128,000 tokens), supporting up to 128,000 completion tokens per response.
Direct Matchups with Codestral 22B
See how Codestral 22B compares against other leading frontier and open-weights models in cost and latency.
Compare Mistral AI vs OpenAI pricing, context limits, and cost per 1M tokens.
Compare Mistral AI vs Groq (Meta) pricing, context limits, and cost per 1M tokens.
Compare Mistral AI vs Google pricing, context limits, and cost per 1M tokens.
Compare Mistral AI vs Google pricing, context limits, and cost per 1M tokens.