A Long-Awaited Debut
After roughly two years of silence, Mira Murati's Thinking Machines Lab has finally released its first artificial intelligence model, dubbed Inkling, now available through the OpenRouter platform. The launch marks a significant moment for the venture founded by the former OpenAI chief technology officer, whose departure and subsequent stealth-mode operation had fueled intense speculation across the AI industry.
The model arrives positioned as one of the strongest open-source options coming out of the Western tech ecosystem, a claim that carries weight given how much of the recent open-weight momentum has come from Chinese labs. For developers and researchers hungry for capable models they can inspect and deploy freely, Inkling represents a fresh entry with serious pedigree behind it.
Two years of silence ended not with a whisper, but with a model that demands to be taken seriously.
Where Inkling Shines
Early testing suggests Inkling delivers genuinely impressive results on certain benchmarks, most notably its performance on the Model Context Protocol, or MCP. That score stands out and signals the model was built with real technical ambition rather than as a rushed proof of concept.
MCP handling matters because it reflects how well a model can interact with external tools and structured data, an increasingly important capability as AI systems move from simple chat interfaces toward more agentic, task-oriented applications. Strong performance there hints at practical utility beyond raw text generation.
The release also underscores a broader shift in the AI landscape, where open-source and open-weight models are steadily closing the gap with proprietary systems from the largest labs.
The Pricing Question
Where the picture grows more complicated is the price-to-performance equation. While Inkling posts strong numbers in specific areas, the cost of running it against the value it delivers is less clear-cut than its headline benchmarks might suggest.
Key considerations for prospective users include:
- Whether the model's standout MCP capabilities justify its operating costs
- How it stacks up against cheaper alternatives for everyday workloads
- The trade-offs between open-source flexibility and raw efficiency
For now, Inkling stands as a credible and closely watched debut, but the ultimate verdict will depend on how the economics play out as more developers put it through its paces in real-world deployments.