OpenAI and Scale are joining forces to assist more corporations profit from fine-tuning our most advanced models.

Corporations expect high performance, steerability, and customization in terms of deploying AI in production. We recently launched fine-tuning for GPT-3.5 Turbo, and can bring fine-tuning to GPT-4 this fall. With fine-tuning, corporations can now securely customize our most advanced models on proprietary data, making our strongest models much more useful. As at all times, data sent out and in of the fine-tuning API is owned by the client and shouldn’t be utilized by OpenAI, or some other organization, to coach other models.

We’re working with Scale as a preferred partner to increase the advantages of our fine-tuning capability given their experience helping enterprises securely and effectively leverage data for AI. Constructing robust enterprise-grade functionality requires rigorous data enrichment and model evaluation. Scale customers can now fine-tune OpenAI models just as they’d through OpenAI, while also benefiting from Scale’s enterprise AI expertise and Data Engine.

Scale has already demonstrated value for purchasers by fine-tuning GPT-3.5 for Brex. Take a look at more details here.