Mira Murati was OpenAI's CTO until last year. Her new AI lab just did the thing OpenAI never would: released a frontier model into the open for anyone to download.
On Wednesday, Thinking Machines Lab released Inkling, a 975-billion parameter multimodal model that handles text, audio, and video. It's open-weight under an Apache 2.0 license, which means any company or developer can grab the weights, run it on their own hardware, and build on top of it without paying anyone a subscription fee.
Inkling isn't small. Running it at full scale takes a cluster of Nvidia's latest B300 chips, which is not something you're spinning up on a laptop. But it uses a mixture-of-experts design that activates only about 41 billion parameters per query, so the actual compute cost per run is dramatically lower than the parameter count suggests. In practice, it's designed to be usable by any serious enterprise without needing hyperscaler infrastructure.
The bigger story here is Murati. Since leaving OpenAI, she's built one of the most closely watched labs in AI. Thinking Machines raised a $2 billion seed round last summer at a $12 billion valuation, the largest seed in history. Backers include Andreessen Horowitz, Nvidia, AMD, and Jane Street.
At the Bloomberg Tech Summit last month, Murati explained the open approach in plain terms. "To get as many people involved as possible, you have to share the knowledge, you have to share tools." Which is a fairly pointed thing to say when the company you used to run built its business around not doing exactly that.
She's been direct about her old employer in other ways too. Reflecting on the Sam Altman ousting that she briefly led OpenAI through in 2023, Murati said the structure of governance in AI decision-making "should not hinge on one person," and that there should be checks and balances. OpenAI has since restructured, but you can read Thinking Machines as a bet that a different kind of AI lab is possible, one that isn't wrapped around a single founder.
The market has been shifting toward her thesis for a while. Enterprises spent the last two years paying premium prices for closed models from OpenAI and Anthropic, and a lot of them are now looking at the bill and wincing. Leonardo Gonzalez, who runs the AI center at consulting firm Trilogy, told Fireworks his team had been "extremely large" accounts at both companies before moving most workloads to open-weight models. His summary of the shift: open models have "reached parity with proprietary models at an order of magnitude less pricing."
That's the wedge Inkling is meant to widen. If a 975B multimodal model with real reasoning ability can be downloaded and hosted anywhere for free, the case for paying premium API rates gets harder every quarter.
Murati is running the mirror image of the company that made her famous. OpenAI built its position by locking down its best models, charging for access, and treating the weights as the entire business. Thinking Machines is trying to prove there's still a great company to build if you give the model away, one that lives in tools, safety layers, and the specialized products built on top. Whether that actually works is the real question hanging over AI in 2026, because if it does, the moat was never the model. It was the story every closed lab told about why the model had to stay closed.
