Mira Murati's Inkling AI Model Launches on OpenRouter
Adoption
Neutral

Mira Murati's Inkling AI Model Launches on OpenRouter

Mira Murati, former OpenAI CTO, released her first AI model from Thinking Machines Lab after two years of development. The Inkling model is now available on OpenRouter and shows strong performance metrics, though cost-efficiency remains mixed.

Jul 26, 2026, 04:07 PM1 min read

Key Takeaways

  • 1## Inkling Model Now Available Mira Murati's Thinking Machines Lab released its debut AI model, Inkling, on OpenRouter following two years without public releases from the lab.
  • 2The model is accessible to users through OpenRouter's API and can be tested directly on the platform.
  • 3## Performance and Pricing Trade-offs Inkling achieved notable scores on the MCP (Massive Comprehensive Performance) benchmark, positioning it competitively against other open-source models available in Western markets.
  • 4However, the model's pricing structure and inference costs create a more nuanced value proposition than raw benchmark numbers suggest.
  • 5The price-to-performance calculation depends on the specific use case and comparison baseline, making straightforward cost-benefit analysis difficult without detailed throughput and latency requirements.

Inkling Model Now Available

Mira Murati's Thinking Machines Lab released its debut AI model, Inkling, on OpenRouter following two years without public releases from the lab. The model is accessible to users through OpenRouter's API and can be tested directly on the platform.

Performance and Pricing Trade-offs

Inkling achieved notable scores on the MCP (Massive Comprehensive Performance) benchmark, positioning it competitively against other open-source models available in Western markets. However, the model's pricing structure and inference costs create a more nuanced value proposition than raw benchmark numbers suggest. The price-to-performance calculation depends on the specific use case and comparison baseline, making straightforward cost-benefit analysis difficult without detailed throughput and latency requirements.

Why It Matters

For Traders

This is not a trading-relevant story; Inkling is an AI model, not a tradeable asset or protocol.

For Investors

Murati's model entry signals continued talent migration from labs to independent ventures; market structure for open-source AI may fragment further.

For Builders

Developers can now evaluate Inkling as an inference option; benchmark scores matter less than latency, cost per token, and downstream application fit.

Sources

Related Articles

Latest News