Thinking Machines Lab releases Inkling 2 Mini open weights
Original title: unsloth/Inkling-2-Mini-GGUF — new model trending #30 on Hugging Face
The 26.2 billion-parameter multimodal model has 3.4 billion active parameters and an Apache 2.0 license.
Unsloth published GGUF quantizations of Inkling 2 Mini, listed as trending on Hugging Face. Thinking Machines Lab describes it as a general-purpose multimodal model that accepts text and images and generates text, released November 20, 2026, under Apache 2.0. It is a 24-layer decoder-only transformer with sparse mixture-of-experts and a looped middle section. The lab says a roughly 20GB mixed NVFP4 and MXFP8 checkpoint can run on listed local hardware.
- Total parameters are 26.2 billion, with 3.4 billion active.
- Three loop passes yield about 5.4 billion effective active parameters per token.
- Each token is routed to 8 of 128 experts.
- The context window is up to 1 million tokens.
- UD-Q4_K_XL is 17.63 GB with 98.0 percent same-top-token versus the Q8 reference.
- Vendors
- Thinking Machines LabUnsloth
- Products
- Inkling 2 Mini
- AI models
- Inkling-2-Mini
From the original article
5 min read at huggingface.coimage-text-to-text · 0 downloads · 76 likes · license apache-2. 0
By unsloth · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.