Thinking Machines Lab releases 975B open-weights model Inkling
Thinking Machines Lab, led by Mira Murati, has released Inkling, a 975B-parameter mixture-of-experts model under Apache 2.0 licensing. The multimodal system trained on 45 trillion tokens represents a significant open-weights entry from a new lab, challenging the concentration of frontier model releases among established players. A smaller 276B variant is forthcoming, signaling a tiered release strategy. The notably sparse model card raises questions about documentation standards in the open-weights ecosystem, even as the release itself expands accessible frontier-scale infrastructure.
Modelwire context
Analyst takeThe sparse model card is doing more work than the headline release. For a 975B-parameter model trained on 45 trillion tokens, the absence of detailed evaluation methodology, training data sourcing, and safety documentation is not a minor gap. It is a signal about where Thinking Machines Lab is in its institutional development, and it sets a precedent for what the open-weights community will accept from credentialed founders.
The documentation gap connects directly to a tension running through recent coverage. The Anthropic coding research covered here on July 16th showed that speed and capability gains can mask downstream costs that only surface later. The same logic applies to model releases: a frontier-scale open-weights drop with thin provenance documentation accelerates adoption while deferring the accountability questions. The open-weights space has been pushing for richer model cards precisely because deployment decisions downstream depend on them.
Watch whether the forthcoming Inkling-Small release ships with materially more complete documentation. If it does, the sparse card on the flagship was a timing constraint. If it does not, this is the lab's working standard, and downstream integrators should price that in.
This analysis is generated by Modelwire’s editorial layer from our archive and the summary above. It is not a substitute for the original reporting. How we write it.
MentionsThinking Machines Lab · Mira Murati · Inkling · Inkling-Small · Simon Willison
Modelwire Editorial
This synthesis and analysis was prepared by the Modelwire editorial team. We use advanced language models to read, ground, and connect the day’s most significant AI developments, providing original strategic context that helps practitioners and leaders stay ahead of the frontier.
Modelwire summarizes, we don’t republish. Simon Willison originally reported this story as “Inkling: Our open-weights model”. The full content lives on simonwillison.net. If you’re a publisher and want a different summarization policy for your work, see our takedown page.