by MistralReleased 1 August 2025
Codestral 2508 is Mistral's code model for fast, repeated edits inside an editor: you send the code before and after a gap and it fills the middle, corrects errors and writes tests. It reads about 256 pages of text or files and answers in text. Pick it for autocomplete where waiting counts.
Reads about 250 pages · writes up to 200 pages
Standard OpenAI fields, grouped by what they control. Anything not listed is accepted and ignored.
Shaping the reply
How long, how varied, where to stop.
max_tokenstemperaturetop_pstopseedfrequency_penaltypresence_penaltyStructured output
Get JSON you can parse instead of prose.
response_formatstructured_outputsTools
Let the model call your functions.
toolstool_choiceOther
Also accepted by this model.
predictionNo. Mistral has retired it. Other Mistral models are on sale.
Codestral 2508 is Mistral's code model for fast, repeated edits inside an editor: you send the code before and after a gap and it fills the middle, corrects errors and writes tests. It reads about 256 pages of text or files and answers in text. Pick it for autocomplete where waiting counts.
Mistral names no languages for Codestral 2508 (batch). Arabic, French and Darja: not stated.
It reads text and PDFs. Up to 256K tokens in one request, about 250 pages. It writes up to 154 000 words at once.
No. Mistral has not published its weights.
Mistral. Released 1 August 2025. It knows the world up to 31 March 2025.
Mistral's flagship speech-to-text model that turns audio into transcripts with word timestamps and speaker labels, with context biasing for domain-specific vocabulary
The Voxtral Mini derivative that only transcribes, taking audio from meetings, voice notes and podcasts and returning the text through a standard transcription request
Mistral's compact speech-to-text model that turns audio into transcripts with word timestamps and speaker labels, and can be steered with context biasing to catch domain terms
Mistral AI's text embedding model from late 2023, whose 1024-dimension vectors capture how passages relate, for semantic search and RAG; Codestral Embed is the version for code
From what its maker and two benchmark sites publish, and where it stands on OpenDunes. Read 26 Oct 2026.
Thinks first No.
Works through the problem before it answers.
Uses your tools
Can call functions your code provides.
Fills your JSON
Answers in the JSON shape you give it.
Searches the web No.
Can look things up online while it answers.
Cheaper on repeats
A prompt it saw a moment ago costs less the second time.
Open weights No.
Read on docs.mistral.ai.
Knows events up to 31 Mar 2025
What it learned stops at this date.
Complete the missing middle of a function in your editor
Write unit tests for an existing module
Fix a failing function from its error message