0

Welcome Inkling 2 Mini, Cutting-edge Performance for your Local Setup

https://huggingface.co/blog/inkling2mini(huggingface.co)
Inkling 2 Mini is a new looped transformer model designed for local AI setups, featuring a multimodal Mixture-of-Experts architecture with approximately 3B active parameters. Its key innovation is a looped design where 12 middle layers are executed three times, allowing it to function like a model with more layers while using fewer physical ones. The model supports a 1M-token context size and accepts both image and text inputs. According to benchmarks, Inkling 2 Mini is state-of-the-art in its class, outperforming models like Gemma 4 31B and Qwen3.8-27B in areas such as coding, reasoning, and vision. Inference support is provided for various frameworks, including Transformers, llama.cpp, and MLX.
0 points•by will22•45 minutes ago

Comments (0)

No comments yet. Be the first to comment!

Have an account? Log in to join the discussion.