Thinking Machines Rolls Out Broad but Efficient Model

Details

Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, has launched Inkling — its first publicly released model. The model is positioned as general-purpose with token efficiency as a defining characteristic, meaning the team explicitly designed around minimizing token consumption rather than maximizing raw capability metrics. The available coverage is sparse on specifics: no benchmark scores, no pricing, no context window size, and no architectural details are disclosed. What is clear is the significance of the release itself — Thinking Machines is among the most closely watched AI startups given Murati's profile and the company's fundraising trajectory since its 2024 launch. The token-efficiency framing positions Inkling to compete in the cost-conscious, mid-tier model segment rather than against frontier capability leaders. This is a pragmatic positioning that acknowledges the crowded top of the market while targeting the large number of production workloads where inference cost and context efficiency matter more than benchmark maximalism. Definitive assessment requires additional coverage: benchmarks, pricing tiers, API availability, and context window details are all absent from current reporting.