Thinking Machines Lab released its first open-weight AI model, Inkling, on Wednesday morning. The system, created by the startup founded by former OpenAI CTO Mira Murati, allows outside developers and companies to download and modify it directly, distinguishing it from the flagship models currently sold by OpenAI, Anthropic, and Google.
Inkling operates as a mixture-of-experts system containing 975 billion parameters total. It draws on approximately 41 billion of those for any given task, a design choice that keeps the model faster and cheaper to run than a full deployment. The system was trained on 45 trillion tokens of text, image, audio, and video data and reasons natively across all three formats.
This launch serves as the company’s first public proof point after 18 months spent building AI infrastructure largely out of public view. Earlier work surfaced in a May research preview of “interaction models,” AI designed to listen, speak, and interrupt rather than wait for user input. The release tests the central bet behind Thinking Machines: that AI organisations adapt for themselves will outperform the one-size-fits-all models the biggest labs currently sell.
The model is designed to provide calibrated answers, including flagging uncertainty rather than guessing. Users can dial thinking effort up or down to trade speed for depth. On one benchmark, the company states Inkling uses a third as many tokens as Nvidia’s Nemotron 3 Ultra to hit the same coding performance. Thinking Machines does not claim Inkling is best-in-class. Briefing materials explicitly state the model is “not the strongest model available today, closed or open.” The focus is instead on well-rounded performance.
That raises a question regarding the target audience beyond the obvious. This is an enterprise product. Thinking Markets is marketing it less as a finished work and more as a starting point for organisations to fine-tune themselves through Tinker, the company’s model-customization platform. OpenAI, Anthropic, and Google have taken a different approach with ChatGPT, Claude, and Gemini, respectively, building general-purpose chatbots first with agentic features layered on top.
A post published by Thinking Machines last week served as the backdrop for this release. The company argued that AI trained centrally by one company and set in stone underperforms AI that organisations shape themselves because much expertise is specific to the people who hold it. The broader idea is that centralized labs sell everyone the same product repeatedly refined by the lab that built it, while enterprises willing to own and customize their own models can wring far more value from them.
An argument gaining steam. In a blog post published Sunday, Microsoft CEO Satya Nadella warned that enterprises using proprietary AI models effectively pay twice. They pay once in subscription costs and again by handing over business knowledge embedded in thousands of prompts and corrections, which can be absorbed into future model versions. Nadella’s company has invested billions in both OpenAI and Anthropic.
Hugging Face CEO Clem Delangue made a similar prediction in conversation with TechCrunch last week. Frontier models will increasingly be reserved for experimentation and high-value tasks, while most production AI work shifts to private or open-source alternatives. That is the exact split Thinking Machines is building around.
The clearest evidence for that argument came recently from a project with Bridgewater Associates, the world’s largest hedge fund. Researchers from both companies took an existing open-source model and trained it further on Bridgewater’s own financial expertise. The result scored 84.7% on financial reasoning tests, beating top proprietary AI models, while costing roughly a fourteenth as much to run. Those results, published jointly in late June, come from the two companies’ own evaluation, not an independent one.
Thinking Machines has also emphasized how quickly it got here. OpenAI took roughly five years and Anthropic roughly three to bring tech to market and show revenue. Thinking Machines says it did the same in about nine months.
Some will wonder whether Inkling was trained on outputs from competitors’ models, a practice known as distillation that has drawn scrutiny industry-wide. The short answer, per the company’s own materials, is partly. Thinking Machines pretrained Inkling from scratch but says it used other open-weight models, including Moonshot AI’s Kimi K2.5, to help generate some of its early post-training data before large-scale reinforcement learning took over. The next model will use fully self-contained post-training instead.
On the cost side, Thinking Machines has been more guarded. It struck a strategic partnership with Nvidia in March to deploy a gigawatt of Vera Rubin computing capacity and says Inkling itself was trained entirely on Nvidia’s GB300 NVL72 systems. The company has not said how it plans to balance that against revenue, which, by most accounts, has not been a primary focus so far. A reported $50 billion fundraising round was said to be coming together last November, which multiple outlets reported had stalled by January. The company has declined to talk about its funding picture since, though Nvidia said it made a “significant investment” in Thinking Machines when the companies announced that March partnership.
A related question is whether Thinking Machines’ spending will ever reach the scale of OpenAI’s or Anthropic’s, or whether its efficiency-driven approach means the economics look different. The company’s bet may be less that it will eventually spend like its larger rivals than that it won’t need to at all. Once weights are public, nothing obligates anyone who downloads them to pay Thinking Machines to run them, unlike the metered access OpenAI and Anthropic sell. Revenue has to come from Tinker, via training, fine-tuning, and a cut of the hosting ecosystem built around it.
Headcount looks more settled. Thinking Machines now employs roughly 200 people, up from levels reported after a wave of departures earlier this year, including two co-founders who left for OpenAI in January.
Thinking Machines does not seem interested in playing up individual moves the way much of the industry does. According to a source inside the company, its culture, by design, favors continuity over reliance on any one personality. It makes sense: it is less of a setback when people change teams if they were never put on a pedestal to begin with. It is also a remarkable thing for a company to insist on, given how much of its own story is still associated with the name of its now-famous co-founder, whether she planned it or not.
