Mira Murati left OpenAI, raised a fortune, and stayed quiet for a year. On July 15 the silence broke: Thinking Machines Lab dropped Inkling, its first public product — and the whole thing is on HuggingFace under Apache 2.0.
What it actually is
Inkling is a foundation model, not an app. 975B total parameters, 41B active (6 of 256 experts per token), 1M context, pretrained on 45T tokens across text, images, audio, and video. Natively multimodal — no bolted-on vision adapter. It’s tuned for coding, tool-calling agents, and RAG, and it beats Nemotron Ultra on the benchmarks they published.
The bet is contrarian. While OpenAI, Anthropic, and Google guard one closed do-everything model, Murati is shipping an open base you own and reshape.
The API angle
You don’t just download weights and hope. Inkling plugs into Tinker, TML’s fine-tuning service, live day zero at 64K and 256K context (50% off for now). Grab the weights, post-train on your data, serve your own version. That’s the pitch: a customizable open foundation, not another rented endpoint.
You Might Also Like
- Thinking Machines Interaction Models Mira Murati Attacks the Turn Based llm
- Cursor Composer 2 Takes on Anthropic and Openai With a 0 50 m Token Coding Model and the Benchmarks Back it up
- Anthropic to Acquire Stainless for 300m Buying the Developer Pipe to Openai and Google
- Meta Model api Muse Spark 1 1 Undercuts Openai and Anthropic at 1 25 4 25 per Million Tokens
- Openai Just Acquired Promptfoo the 86m ai Security Startup Used by 25 of Fortune 500
Top AI Product 
Leave a comment