Last updated: 16/04/2025

MistralOfficial Docs

Codestral

codestral-2501

active

Codestral

Codestral is Mistral's cutting-edge language model for coding tasks, specializing in low-latency, high-frequency operations like fill-in-the-middle (FIM), code correction, and test generation. Released in January 2025, this powerful model excels at a wide range of text-based tasks.

Supports a 262K token context window. Handles Text, Image, Video, Audio, Transcription, Text-to-Speech inputs and outputs. Supports fine-tuning for custom applications. Supports tool use for advanced automation. Capable of generating structured output formats.

Additional Information

Notes

Context length: 262144 tokens. Available via API as 'codestral-latest'. Has aliases: codestral-latest, codestral-2412, codestral-2411-rc5.

Model Timeline

Launch Date

1/1/2025

Last Updated

1/1/2025

Capabilities

Text

Input Pricing

$-/ KTok

Context: 262,144 tokens

Output Pricing

$-/ KTok

Max tokens: 262,144

Embeddings

Embeddings Pricing

$0.10/1k tokens

Additional Model Information

Tool Use

Yes

Structured Output

Yes

Reasoning

Yes

Flatten your repo for AI in seconds

Flatten repos. Prompt faster. One click → one GPT-ready file

Free Online & Desktop