Thinking Machines releases Inkling 2 Mini, an open looped model that runs on a laptop
Thinking Machines Lab has released Inkling 2 Mini, an open-weight model small enough to run on a laptop, the company said on 20 November. It follows the larger Inkling 2 Small released earlier this month.
Inkling 2 Mini is a mixture-of-experts model with 26.4 billion total and 3.4 billion active parameters. It is a "looped" transformer: its middle layers run three times for every token, which the company says lets it apply more reasoning and reach the right answer with fewer tokens, especially on agentic coding. Most weights were trained and run natively in NVFP4, giving a memory footprint of about 20GB.
The company reports that the model leads every open-weight model under 100 billion total parameters on benchmarks such as DeepSWE v1.1 (48.2%) and Vals Index v2.2 (46.5%), while trailing Qwen3.8-27B on some others. Looping does not help it store more facts, so it relies on search and tools for knowledge-heavy tasks. Hugging Face, which supports the model in transformers, llama.cpp and MLX, calls it "at least comparable" with Qwen 3.8 27B in its own vibe tests, adding that more validation is needed.
Because looping lets a model reason longer without writing its reasoning down, Thinking Machines also released a research variant with a variable number of loops to study how that affects monitoring. In its tests, the chain of thought grew shorter with more loops but stayed legible, and the model did not score higher on tests of agentic sabotage. Inkling 2 Mini is available on Hugging Face and on the company's Tinker platform.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like AI Programming Weekly or Daily AI Brief.
Also, consider following us on social media:
AI Programming Weekly
Weekly news about AI tools for software engineers, AI enabled IDE's and much more.
Industry analysis
2025 Global Business Services Agenda: Gen AI Takes Center Stage
This industry analysis by The Hackett Group explores the transformative impact of generative artificial intelligence (Gen AI) on global business services (GBS) in 2025. The study highlights the shift from exploration to acceleration of Gen AI initiatives, with 89% of executives advancing these projects to improve customer satisfaction, innovate products, and reduce costs. The report also discusses the challenges and strategies for successful Gen AI adoption, emphasizing the need for a technology-enabled operating model and the importance of reskilling the workforce.
Read moreYou may also like
Thinking Machines Lab seeks $5 billion at a $40 billion valuation, with Accel to lead, report says
Liquid AI Releases LFM3, Six Open-Weight Models for On-Device AI
MBZUAI's Institute of Foundation Models releases K2-Horizon, six fully open models up to 375B parameters
Agnes AI releases open-weight Agnes 2.5 Pro Theta, a 188B multimodal model
Ricursive releases Ripple 65B Preview coding model under Apache 2.0
Daily AI Brief: the AI news that matters, in your inbox.