Mistral challenges OpenAI with new code model Codestral Embed
Mistral AI has introduced its first specialised embedding model for code, Codestral Embed. The model outperforms competitors like OpenAI in code retrieval at a lower price point.
What happened?
French AI developer Mistral AI has launched its first code-specific embedding model, Codestral Embed. The model is designed to convert source code and technical data into numerical vectors, optimising the performance of Retrieval-Augmented Generation (RAG) systems. According to the company's internal testing, Codestral Embed outperforms rival options such as OpenAI Text Embedding 3 Large, Cohere Embed v4.0, and Voyage Code 3 in searches across real-world codebases.
Key facts
| Modellnamn | Codestral Embed |
|---|---|
| Pris per miljon tokens | 0,15 USD |
| Lanseringsdatum | 28 maj 2025 |
”Super excited to announce the official release of @MistralAI Codestral Embed, our first code-specific embedding model.”
Why it matters
Code embeddings are critical for enabling AI systems to rapidly identify and retrieve the correct source code segments from massive repositories. By offering high performance at a price of 0.15 USD per million tokens, coupled with support for flexible vector dimensions, Mistral makes it significantly more affordable and efficient for companies to build advanced code assistants.
Who is affected?
The news primarily concerns software developers, data engineers, and companies building AI-assisted coding tools or internal search systems for large source code repositories. Developers seeking more cost-effective and accurate embedding models for RAG architectures now have a new European alternative.
Impact on the EU
Codestral Embed is available globally via Mistral AI's API for all developers, including those within the EU. As Mistral AI is a French company, its data and model handling adhere to European data protection regulations, such as the GDPR and the upcoming requirements of the EU AI Act.
What else you should know
The model is part of Mistral's broader initiative to provide code-specific AI tools under the Codestral umbrella. By offering support for lower dimensions, such as 256, and int8-quantisation, Mistral also addresses the high infrastructural storage costs often associated with large-scale RAG systems.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Är modellen tillgänglig i EU?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
AI-verktyg i artikeln
Topics
Get similar news straight to your inbox
The reader's room
Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.
Sign in to submit a comment or question.
Read the article through your role
- Assess technical risk: model choice, vendor lock-in, data flow and running cost.
- Update the architecture doc if new APIs or regulations touch production.
- Ensure observability + rollback plan before rolling out to production.
Generated angle — not editorial analysis of "Mistral challenges OpenAI with new code model Codestral Embe"