88Z AI, 22 November — Thinking Machines has launched Inkling 2 Mini, an open-source model using a looped Transformer architecture that runs specific layers multiple times per token to enhance internal reasoning. It delivers high performance at a significantly lower cost and can run on laptops.
At a glance
- Inkling 2 Mini uses a looped Transformer architecture to enhance internal reasoning by repeating layer execution
- Matches Qwen3.8 27B performance at roughly one-sixth the cost and supports local laptop deployment
Architectural Innovation
Thinking Machines has officially released Inkling 2 Mini, an open-source model featuring a unique looped Transformer architecture. Twelve of its 24 layers are executed three times per token, allowing the model to engage in extended internal reasoning without generating additional text output. This approach enables it to solve complex problems using fewer words.
Performance and Cost Efficiency
According to the Artificial Analysis Intelligence Index, Inkling 2 Mini matches the performance of Qwen3.8 27B while operating at approximately one-sixth of the cost. The model is optimized to run locally on standard laptops, providing developers and users with accessible high-performance inference capabilities.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- the-decoder.comSource published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.