Former OpenAI CTO Mira Murati's startup drops its first model with nearly a trillion parameters, a million-token context window, and a clear message: open beats closed.
Mira Murati left OpenAI in September 2024. Less than two years later, her startup just shipped a model that makes most open-source alternatives look like science fair projects.
Thinking Machines Lab launched Inkling on July 15, a multimodal AI model packing 975 billion total parameters. It processes text, images, audio, and video. It supports a context window of up to 1 million tokens. And it’s available for free download on Hugging Face.
What Inkling actually is #
Inkling uses a Mixture-of-Experts (MoE) architecture. Think of it like a massive office building where 975 billion employees work, but only 41 billion show up for any given task. The rest stay home. This design means you get the knowledge depth of a near-trillion parameter model without needing the compute budget of a small nation to run it.
The model was trained on 45 trillion tokens across multiple modalities. To put that in perspective, GPT-3 was trained on roughly 300 billion tokens. We’re talking about a training dataset roughly 150 times that size.
One of the more interesting design choices is what Thinking Machines calls controllable “thinking effort.” In English: you can dial up or down how hard the model reasons through a problem. Simple customer service query? Low effort, fast response, cheaper compute. Complex multi-step analysis? Crank it up. This kind of efficiency toggle matters enormously when you’re paying per token at scale.
Alongside the flagship model, the company also released Inkling-Small, a lighter variant with 12 billion active parameters. It’s designed for teams that need capable AI but don’t have access to enterprise-grade GPU clusters.
Both models are available on Hugging Face, with a variant specifically optimized for NVIDIA’s Blackwell hardware.
Fine-tuning is the real product #
Thinking Machines built a fine-tuning platform called Tinker that lets developers adapt Inkling for specific tasks without the PhD-level expertise that fine-tuning typically demands. The pitch is straightforward: take a powerful base model, shape it to your exact use case, and deploy it at a fraction of what a proprietary API would cost over time.
Inkling’s open-weights approach means developers can download the model, run it on their own infrastructure, and fine-tune it however they want.
The money behind the mission #
Murati founded Thinking Machines Lab in February 2025, roughly five months after her departure from OpenAI. The company raised a $2 billion seed round at a valuation between $10 billion and $12 billion, with backing from investors including Andreessen Horowitz and Nvidia.
Murati wasn’t just any OpenAI employee. She served as CTO and briefly as interim CEO during the Sam Altman boardroom drama in late 2023.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our