Mira Murati, the former chief technology officer at OpenAI who departed amid last year's boardroom drama, has launched a new frontier AI model through her startup Thinking Machines Labβ€”and this one is genuinely open in ways that Sam Altman's company hasn't been willing to match.

The Model That OpenAI Wouldn't Build

The new model, released under permissive licensing terms, includes full access to weights and architectural details. This stands in sharp contrast to OpenAI's trajectory under Altman, which has increasingly moved toward closed deployment even as the company positioned itself as a leader in AI safety and transparency. Murati reportedly raised $2 billion for Thinking Machines Lab, giving her serious resources to back up the technical claims.

Why This Matters for Developers

For the developer community, this represents one of the most capable open-weight models available at the frontier level. The ability to run, modify, and study a top-tier model without API restrictions or usage caps addresses long-standing frustrations with proprietary alternatives. Researchers can inspect training methodologies, fine-tune without licensing negotiations, and deploy in environments where data sovereignty is non-negotiable.

Industry Implications

The move puts additional pressure on OpenAI's narrative around safety and responsible release. Murati's team appears to be betting that openness and capability aren't mutually exclusiveβ€”and if the benchmarks hold up, it could reshape how frontier labs think about model deployment going forward.

Key Takeaways

  • Mira Murati's Thinking Machines Lab releases a fully open frontier model with public weights
  • The $2 billion-backed startup positions this as a direct challenge to closed AI development approaches
  • Technical capabilities reportedly compete with GPT-4o class models while maintaining openness commitments

The Bottom Line

Murati doing what Altman won'tβ€”actually releasing frontier technology openlyβ€”is the kind of move that forces the entire industry to confront whether 'safety concerns' are legitimate or just competitive moat-building dressed up in ethics language. Either way, developers win.