Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, released its first in-house AI model Wednesday morning, called Inkling. Unlike the flagship models from OpenAI, Anthropic, or Google, Inkling is open-weight, meaning outside developers and companies can download it and modify it directly. This move positions Thinking Machines as a challenger to the dominance of closed-source AI models.

Inkling is a mixture-of-experts system with 975 billion total parameters, though it only draws on a fraction of that — about 41 billion — for any given task. This design keeps very large models efficient while maintaining performance. The company trained Inkling on 45 trillion tokens of text, image, audio, and video, and the model reasons natively across all four modalities.

The release of Inkling comes as Microsoft CEO Satya Nadella warned that enterprises using proprietary AI models effectively pay the cost of innovation to a few companies. Thinking Machines now employs roughly 200 people, a significant increase from earlier this year after a wave of departures, including two co-founders who left for OpenAI.

The open nature of Inkling may shift the balance of power in the AI industry, reducing reliance on a few dominant players. However, it also raises questions about governance, security, and the potential for fragmentation in model development. Companies may face trade-offs between customization and the risks of managing open-source models internally.

Thinking Machines' move with Inkling signals a broader trend toward openness in AI, though the long-term impact remains uncertain. The company's approach may influence how other startups and established players navigate the evolving landscape of AI development and deployment.