Mistral AI Free tierOpen weights

Codestral 25.08

Codestral 25.08 is a Codestral model from Mistral AI, released in 2025-07. It costs $0.30 / $0.90 per 1M, has a 256k-token context window, and is best for coding, fill-in-the-middle, cheap-coding. Last verified 2026-04-19.

Try Codestral 25.08 Docs Take the picker quiz

Spec sheet

Pricing

Input
$0.3 / 1M
Output
$0.9 / 1M
Free tier
La Plateforme

Context & speed

Context window
256k tokens
Max output
8k tokens
Throughput
~170 tok/s
Time to first token
~320 ms
Speed tier
fast

Capabilities

Tool use
Yes
Structured output
Yes
Prompt caching
No
Extended thinking
No
Vision input
No
Audio in / out
No
Fine-tuning
Yes

Deployment

Open weights
Yes
On-prem
Yes
HIPAA eligible
Yes
Zero retention
Yes
Regions
eu, us

Estimated monthly cost

Assumes typical token shape: 2k input, 600 output per call. Prompt caching is excluded from these figures.

10k calls/mo
$11.40
per month
100k calls/mo
$114.00
per month
1M calls/mo
$1.1k
per month

When to use Codestral 25.08

Sweet spot

  • coding
  • fill in the middle
  • cheap coding
  • ide autocomplete

Known trade-offs

  • coding-only — not general purpose

Best use cases

Best LLM for Coding assistant

Best price/performance for fill-in-the-middle and IDE completion

Works with

Mistral SDKContinue.devCursorOpenRouterOllamavLLM

Compare Codestral 25.08 to other models

vs DeepSeek V3.2vs Gemini 3.1 Flash-Litevs Gemini 3 Flashvs Qwen 3.6 Plusvs GLM-5.1vs o4-mini

FAQ — Codestral 25.08

How much does Codestral 25.08 cost?

Codestral 25.08 costs $0.30 / $0.90 per 1M tokens on the Mistral AI API. This model does not currently support prompt caching, so list price is the full cost.

What is the context window of Codestral 25.08?

Codestral 25.08 has a 256k-token context window with up to 8k tokens of output. That's enough for long reports, extended chat histories, or structured document analysis.

Does Codestral 25.08 have a free tier?

Yes — Free tier for prototyping with rate limits. Start at https://console.mistral.ai.

Is Codestral 25.08 HIPAA / EU / on-prem friendly?

Codestral 25.08 is HIPAA-eligible, available in EU regions, and offers open weights for self-hosting. Zero data retention is available for enterprise customers.

What is Codestral 25.08 best for?

Codestral 25.08 is best for coding, fill in the middle, cheap coding, ide autocomplete. Trade-offs to be aware of: coding-only — not general purpose.

Which tools and SDKs work with Codestral 25.08?

Codestral 25.08 integrates with Mistral SDK, Continue.dev, Cursor, OpenRouter, Ollama, vLLM. Most major AI frameworks support it either natively or through OpenAI-compatible endpoints.