FrontierJuly 31, 2026via The Decoder
Thinking Machines bets on efficiency over size with its second model, Inkling Small
Why it matters
Open-weight reasoning models are becoming more efficient; Inkling Small's sub-linear scaling relative to its predecessor signals a shift in how frontier labs optimize for cost and capability tradeoffs. Practitioners evaluating reasoning models now have a new efficiency benchmark to consider.
Key signals
- Thinking Machines released Inkling Small
- Inkling Small is less than one-third the size of the original Inkling model
- Inkling Small beats original Inkling on multiple coding and reasoning benchmarks
- Open-weights reasoning model
- Lab leadership: Mira Murati (former OpenAI CTO)
- Published 31 July 2026
The hook
Mira Murati's Thinking Machines just shipped a smaller model that outperforms its predecessor on reasoning and code — a direct bet against the scaling narrative.
Thinking Machines, the AI lab from former OpenAI CTO Mira Murati, has released Inkling Small. The open-weights reasoning model is less than a third the size of Inkling but beats it on several coding and reasoning benchmarks.