Darius Varu
January 14, 2025 01:51
Mistral AI has released Codestral 25.01, an advanced coding model that improves code generation speed and efficiency. Now available in leading AI code assistants.
In a significant development for software developers, Mistral AI announces the release of Codestral 25.01, an upgraded version of its state-of-the-art coding model. According to Mistral AI, this new version promises to double the speed of code generation, offering a more efficient architecture and improved tokenizer.
Advances in AI Code Generation
Codestral has changed the game for AI code generation by supporting over 80 programming languages and improving productivity through tasks such as fill-in-the-middle (FIM), code modification, and test generation. The newly released Codestral 25.01 builds on this legacy and sets a new benchmark in the realm of AI coding.
Benchmark performance
Codestral 25.01 outperforms several other leading models in a variety of benchmarks. It has demonstrated excellent performance on HumanEval and FIM tasks in tests using popular programming languages such as Python, Java, and JavaScript. The model’s efficiency is highlighted by its 256k context length and high accuracy across a variety of coding tasks.
Global accessibility and integration
Starting today, Codestral 25.01 is available to developers worldwide through IDE and plugin partners. For enterprises that require data and model residency, models can be deployed locally on premises or through a virtual private cloud. Mistral AI is facilitating integration through platforms such as Continue, with demos available to VS Code and JetBrains users.
Industry Impact
Ty Dunn, co-founder of Continue, emphasized the importance of Codestral 25.01 in improving code completion tasks, a critical area for AI code assistants. The ability to provide accurate and fast proposals will make this model the preferred choice for developers aiming for efficient software development.
future prospects
Developers interested in integrating with Codestral’s API can access it through Vertex AI on Google Cloud, with plans for broader availability on platforms such as Azure AI Foundry and Amazon Bedrock. This model’s top ranking on the LMsys Co-Pilot Stadium leaderboard highlights its potential to redefine AI-based coding.
Image source: Shutterstock