Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has unveiled its first product, Tinker.

The launch of Tinker ends months of speculation on what Murati and her team were working on. In July, Thinking Machines Lab had also attracted a whopping $2 billion in a seed funding round led by Andreessen Horowitz at a $12 valuation even before making the product public.

In a blogpost, Thinking Machines Lab said that Tinker is a flexible API for fine-tuning language models. The new product helps researchers and hackers to experiment with models by giving them control over the algorithms and data, it explained. Thinking Machines Lab will handle the complexity of distributed training, it added.

The new product presently allows users to fine-tune two open source models -- Meta’s Llama and Alibaba’s Qwen. 

It said that groups at Princeton, Stanford, Berkeley, and Redwood Research have already been using Tinker

For instance, the Rotskoff Chemistry group at Stanford fine-tuned a model to complete chemistry reasoning tasks using Tinker, it pointed.

Currently, the product is in private beta for researchers and developers.