Mistral AI has been a leader in the code generation field and its latest model, Codestral 25.01, continues to push boundaries. This upgraded iteration promises even greater speed and capability, setting a new benchmark for coding models in its class. Described as lightweight, fast, and proficient across more than 80 programming languages, the original Codestral proved invaluable for developers tackling tasks like fill-in-the-middle (FIM), code correction, and test generation. The model quickly gained recognition for its ability to enhance productivity, serving as a reliable coding companion for developers worldwide. With a refined architecture and an upgraded tokeniser, Codestral 25.01 delivers code generation and completion abilities that are double the speed of its predecessor. For FIM use cases, Codestral 25.01 now leads the pack and is the definitive SOTA model in its weight class.

Codestral 25.01 Performance Benchmarks

To showcase the impressive capabilities of Codestral 25.01, Mistral AI has evaluated its performance against leading sub-100 billion parameter coding models, renowned for their excellence in FIM tasks. The results are a testament to the significant strides made with this latest model, as highlighted by key coding metrics:

Tailored for Developers

Codestral 25.01's improvements are a direct result of prioritizing practical applications and incorporating developer insights. Whether utilized through integrated development environment (IDE) partners or embedded within enterprise workflows, the model is built for seamless real-world integration. Developers working with IDEs that support Codestral 25.01 will immediately benefit from faster and more accurate code completion. Combined with its exceptional performance in FIM tasks and other coding challenges, Codestral 25.01 solidifies its position as an essential tool for agile software development.

"Code completion forms a significant part of the workload for AI coding assistants, necessitating models that excel at fill-in-the-middle (FIM) tasks," says Ty Dunn, co-founder of Continue, a company specializing in enterprise-grade implementations of Mistral's models. "Mistral AI's latest model delivers more accurate suggestions at much higher speeds, which is essential for efficient and precise software development. This is why we recommend Codestral as the go-to autocomplete solution for developers."

Available Across Platforms

Mistral AI ensures that Codestral 25.01 reaches developers through various channels. The model is accessible via major APIs, including Google Cloud's Vertex AI, currently in private preview on Azure AI Foundry, and will soon be available on Amazon Bedrock. This versatility allows both developers and enterprises to seamlessly integrate Codestral 25.01 into their workflows. In addition, Mistral provides open access to Codestral's documentation and API through its la Plateforme portal, enabling developers to explore integrations and tailor the model's capabilities to their unique requirements. For hands-on use, Codestral 25.01 is also available as IDE plugins for platforms such as VS Code and JetBrains, making it easy for developers to experience its benefits firsthand. With Codestral 25.01, Mistral AI reinforces its leadership in the coding model revolution. Combining efficiency, accuracy, and scalability, the model paves the way for transformative innovations that will redefine the future of software development.