Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, released its first proprietary AI model, Inkling, on Wednesday. Unlike the flagship models from OpenAI, Anthropic, or Google, Inkling is open-weight, meaning outside developers and companies can download and modify it directly. The model is a mixture-of-experts system with 975 billion total parameters, though it only uses about 41 billion for any given task, a design that keeps large models faster and cheaper to run.
Inkling's Design and Performance
Inkling was trained on 45 trillion tokens of text, image, audio, and video, and reasons natively across all three modalities, according to the company. It is designed to give calibrated answers, including flagging uncertainty rather than guessing, and allows users to dial "thinking effort" up or down to trade for speed. On one benchmark, the company claims Inkling uses a third as many tokens as Nvidia's Nemotron 3 Ultra to achieve the same coding performance. However, Thinking Machines explicitly states that Inkling is "not the strongest model available today, closed or open," instead aiming for well-rounded performance.
Enterprise Focus and Customization
Thinking Machines is marketing Inkling less as a finished product than as a starting point for organizations to fine-tune through its model-customization platform, Tinker. This approach shifts responsibility for safety and customization to customers, who need serious machine learning talent to fine-tune effectively. The company argues that AI trained centrally and set in stone underperforms AI that organizations shape themselves, because expertise is often specific to the people who hold it. This argument is gaining traction: Microsoft CEO Satya Nadella recently warned that enterprises using proprietary AI models effectively pay twice—once in subscription costs and again by handing over business knowledge embedded in prompts and corrections. Hugging Face CEO Clem Delangue similarly predicted that frontier models will be reserved for experimentation while most production AI shifts to private or open-source alternatives.
Evidence and Speed to Market
The clearest evidence for Thinking Machines' argument came from a recent 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 financial expertise, achieving 84.7% on financial reasoning tests—beating top proprietary AI models—while costing roughly a fourteenth as much to run, though these results come from the companies' own evaluation. Thinking Machines emphasizes its speed: OpenAI took roughly five years to bring tech to market and show revenue, and Anthropic roughly three, while Thinking Machines says it did the same in about nine months.
Training Data and Economics
Regarding whether Inkling was trained on outputs from competitors' models, the company says it pretrained Inkling from scratch but used other open-weight models, including Moonshot AI's Kimi K2.5, to help generate some early post-training data before large-scale reinforcement learning took over. The next model will use fully self-contained post-training. On the cost side, Thinking Machines struck a strategic partnership with Nvidia in March to deploy a gigawatt of Vera Rubin computing capacity, and Inkling was trained entirely on Nvidia's GB300 NVL72 systems. However, the company hasn't detailed how it plans to balance that against revenue, which hasn't been a primary focus so far. A reported $50 billion fundraising round was said to be coming together last November but stalled by January; the company has declined to discuss its funding picture since, though Nvidia said it made a "significant investment" in Thinking Machines.
Business Model and Culture
Thinking Machines' bet may be that it won't need to spend like its larger rivals because once weights are public, nothing obligates anyone who downloads them to pay Thinking Machines to run them. Revenue must come from Tinker—training, fine-tuning, and a cut of the hosting ecosystem built around it. Headcount now stands at 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. According to a source inside the company, its culture favors continuity over reliance on any one personality, making it less of a setback when people change teams. This is notable given how much of the company's story is still associated with its famous co-founder.