1 hour ago
31
Mira Murati left OpenAI. Then she raised $2 billion. Now she’s giving away the goods.
Thinking Machines Lab, the AI startup founded by OpenAI’s former CTO, launched its first open-weight model on July 15. It’s called Inkling, and it’s not small. The model packs 975 billion total parameters in a Mixture-of-Experts architecture, with 41 billion active at any given time. It processes text, audio, and video, and supports a context window of up to 1 million tokens.
Murati is releasing it under an Apache 2.0 license, meaning developers can download, modify, and deploy it with virtually no strings attached.
What Inkling actually is
The Mixture-of-Experts approach is worth unpacking. In English: instead of firing up all 975 billion parameters for every query, the model activates only a relevant subset, roughly 41 billion parameters, depending on the task. It keeps compute costs manageable while preserving the depth of a much larger model.
Alongside the flagship model, Thinking Machines also revealed Inkling-Small, a lighter variant with 12 billion active parameters.
Both models are available for immediate download and integrate with platforms like Databricks and Hugging Face.
The bigger play against closed AI
Murati founded Thinking Machines Lab in February 2025, just months after departing OpenAI. The company raised $2 billion in seed funding at a valuation of roughly $10 billion. The contrast with competitors is deliberate. OpenAI and Anthropic keep their most powerful models locked behind API access, controlling how customers interact with the technology. Thinking Machines is betting that enterprises want something different: models they can customize, fine-tune, and run on their own infrastructure without handing their data to a third party.
Back on May 11, the company released TML-Interaction-Small, a model focused on real-time interactions with 276 billion total parameters and 12 billion active. Inkling is the main event.
What this means for investors and the AI market
There was no immediate market reaction to the Inkling launch. Thinking Machines is still private.
Meta has been the dominant player in open-weight AI with its Llama series. Mistral has carved out a niche in Europe. Now Murati, a high-profile departure from OpenAI, is staking a claim with a model at 975 billion parameters under Apache 2.0 licensing.
The risk, as always with open models, is monetization. Giving away your most impressive technology under a permissive license is a bold strategy when you’ve raised $2 billion. Thinking Machines will need to prove that customization, enterprise support, and platform integrations can generate enough revenue to justify the $10 billion valuation.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·