Sponsored by Byond Boundrys Consulting - Empowering Ideas, Delivering Results
Mistral AI Intermediate

Codestral 25.08

Codestral 25.08 is Mistral’s low-latency coding model for code completion, fill-in-the-middle and IDE assistants, with a 128K context and $0.30/M input pricing.

Coding ModelsText Paid
In plain English

What is this model and why does it matter?

Codestral 25.08 is Mistral's current coding model for low-latency code completion, fill-in-the-middle, code generation and developer workflows.

Code completionFill-in-the-middleCode generationIDE assistantsDeveloper automation
Model overview

Codestral 25.08: features, use cases and important details

Codestral 25.08 is a current AI model verified from first-party Mistral AI sources.

Codestral 25.08 verified specifications

Codestral 25.08 is Mistral’s current coding model for low-latency code completion, fill-in-the-middle, code generation and developer workflows. Its verified context or usage limit is 128000 tokens.

Codestral 25.08 pricing and access

Codestral 25.08 costs $0.30 per 1M input tokens, $0.03 per 1M cached input tokens and $0.90 per 1M output tokens.

Codestral 25.08 best uses

Code completion, Fill-in-the-middle, Code generation, IDE assistants, Developer automation.

Codestral 25.08 limitations

Generated code needs tests, No open weights for this current endpoint, Not ideal for broad non-coding tasks.

Get started

How to use this model

  1. Create a Mistral API key.
  2. Call codestral-2508.
  3. Use FIM for inline completion or chat for coding tasks.
  4. Use functions or structured outputs when needed.
  5. Test generated code.
Copy and try

Example prompts

  • Complete the missing function in this file.
  • Generate tests for this module.
  • Fill the code between this prefix and suffix.
Capabilities

What it can do

  • 128K context
  • FIM
  • Code completion
  • Function calling
  • Structured outputs
  • Predicted outputs
Best for

Practical use cases

  • IDE completion
  • Coding assistants
  • Test generation
  • Code correction
  • Developer tools
Pricing

What does it cost?

Codestral 25.08 costs $0.30 per 1M input tokens, $0.03 per 1M cached input tokens and $0.90 per 1M output tokens.

Input$0.30 / 1M tokens
Output$0.90 / 1M tokens
Simple summaryVery low code-model pricing at $0.30/M input and $0.90/M output.

What stands out

  • Low price
  • Purpose-built FIM
  • 128K context
  • Current GA endpoint

Things to consider

  • Specialized rather than general-purpose
  • Closed hosted model
Limitations

Important restrictions and trade-offs

  • Generated code needs tests
  • No open weights for this current endpoint
  • Not ideal for broad non-coding tasks
SimplifyAITools verdict

Our editorial take

A useful high-demand coding page that clearly distinguishes the current 25.08 Codestral endpoint from the older 24.05 model already on the site.

References

Primary sources

  1. Open source 1 ↗
  2. Open source 2 ↗
  3. Open source 3 ↗
Keep exploring

Related AI models