Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Hugging Face ships Inkling 2 Mini, a looped-transformer MoE for local setups

AI summary

Hugging Face announced that Thinking Machines Lab's Inkling 2 Mini is available, describing it as a multimodal mixture-of-experts model with about 26B physical total parameters and 3B active parameters per token. Its 24 decoder layers include 12 middle layers that run three times in a loop, giving it the effect of 48 logical layer visits, and it supports a 1M-token context with image and text inputs. Weights ship in mixed precision with NVFP4 and MXFP8 expert weights, a DSpark drafter is included, and support is available in transformers, Llama App, llama.cpp, MLX and SGLang with day-zero inference. Hugging Face says in its informal vibe tests the model was at least comparable with Qwen 3.8 27B, while noting that more validation is needed.

In-depth insight

A looped-transformer multimodal MoE at this size gives local and datacenter users another option for long-context, vision-capable inference, though the quality comparison against Qwen 3.8 27B remains informal and unvalidated.

Today we are releasing Inkling 2 Mini, a fast, powerful model that’s small enough to run on your laptop.

Inkling 2 Mini is a looped transformer, running the middle layers of the model three times for every token. This allows the model to apply more reasoning and reach the right answer with fewer tokens, especially on tasks such as agentic coding that depend on reasoning quality. Inkling 2 Mini leads every open-weights model with less than 100B total parameters on benchmarks such as DeepSWE v1.1 (48.2%) and Vals Index v2.2 (46.5%). It runs faster than similarly performing models as well.

DeepSWE v1.1

SWE-bench Pro v2 Hard (51 tasks)

Reported coding scores for models under 100B total parameters. The green line is the frontier with Inkling 2 Mini, the dashed line the frontier without it, and the shading the ground it adds. Decoding speed is a theoretical ceiling at 614 GB/s with BF16 weights and batch size one, including repeated reads across loops. Evaluation setups vary by model.

Inkling 2 Mini is a Mixture-of-Experts model with 26.4B total parameters, 3.4B active, and 5.4B weight activations per token to account for looping. Looping allows the model to reason longer without writing its reasoning down, which matters for monitoring: our ability to oversee and understand what the model does by reading its chain of thought (CoT).

Some source content is omitted. Continue reading on the original site.

#Inkling 2 Mini#Thinking Machines Lab#Hugging Face#Qwen 3.8 27B#SGLang#Llama App#llama.cpp#MLX#transformers#DSpark#Nanbeige

Original · 21 min readRead the original
Group chat

No comments yet. Start the conversation.