Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, has released its debut foundation model, according to VentureBeat. Named Inkling, the model marks the company's first release trained entirely from scratch, and its full weights are available for anyone to download, fine-tune, and deploy.

What sets Inkling apart is its combination of scale and accessibility. The model is built as a mixture-of-experts transformer with a massive parameter count, yet only a fraction of those parameters activate for any given task, keeping inference costs manageable. It also natively handles text, image, and audio inputs, and supports a context window large enough to process lengthy documents or extended conversations in a single pass.

For enterprises, the appeal lies in control. Rather than depending solely on closed, proprietary APIs, organizations can run Inkling on their own infrastructure or within private clouds, tailoring it to specific compliance, cost, or customization needs. This kind of open-weights release lowers the barrier for companies wanting to build differentiated AI products without starting from zero.

As open foundation models continue to mature alongside their closed counterparts, businesses have more genuine choice in how they architect their AI strategies. At Generative AI Solutions, we track releases like Inkling closely to help clients choose the right foundation for their own AI products.