Mistral introduces new services and SDK for customers to customize its models

French AI startup Mistral is shaking up the AI customization game with its new offerings for developers and enterprises. Mistral is now providing new AI model customization options, including paid plans, to allow users to fine-tune its generative models for specific use cases.

One of the key features Mistral is introducing is self-service customization through its software development kit (SDK), Mistral-Finetune. This SDK allows users to fine-tune Mistral’s models on various platforms, from workstations to small datacenter nodes. The SDK is optimized for multi-GPU setups but can also work with a single Nvidia A100 or H100 GPU for smaller models like Mistral 7B.

For those who prefer a more managed solution, Mistral is offering fine-tuning services through its API. Currently compatible with Mistral Small and Mistral 7B, these services will expand to support more models in the future. Additionally, Mistral is rolling out custom training services for select customers, allowing organizations to fine-tune any Mistral model for their specific applications using their own data.

With Mistral seeking to raise $600 million at a $6 billion valuation, the company is clearly aiming to increase its revenue in the face of growing competition in the generative AI space. Since unveiling its first generative model in September 2023, Mistral has continued to innovate with new models and paid APIs. While the company has not disclosed user numbers or revenue figures, its latest customization options are sure to attract attention from developers and enterprises looking to optimize their AI models for specific needs.