Issue #35: Mistral announces new code generation model

Introducing the new Codestral model

Welcome to Issue #35 of One Minute AI, your daily AI news companion. This issue discusses the new announcement from Mistral AI.

Source: Solen Feyissa on Unsplash

---

Introducing Mistral’s new code generation model

Mistral AI has introduced Codestral, a groundbreaking generative AI model tailored for code generation. Codestral is designed to streamline coding tasks such as completing functions, writing tests, and generating documentation. It supports over 80 programming languages, including Python, Java, C, C++, and Bas, making it a versatile tool for developers across different platforms.

The model is powered by a 22 billion parameter architecture and features a 32,000 token context window, enabling it to handle complex code generation tasks efficiently. This extensive context window allows Codestral to understand and generate longer and more contextually accurate code snippets compared to previous models.

With its larger context window of 32k (compared to 4k, 8k, or 16k for competitors), Codestral outperforms all other models in RepoBench, a long-range eval for code generation (source: Mistral)

You can download the model from Hugging Face or access it via codestral.mistral.ai for free during an 8-week beta. It is also integrated into platforms like LlamaIndex, LangChain, VSCode, and JetBrains, as well as in Le Chat for conversational use.

---

Want to help?

If you liked this issue, help spread the word and share One Minute AI with your peers and community.

You can also share feedback with us, as well as news from the AI world that you’d like to see featured by joining our chat on Substack.