GeethanTechGeethanTechTech, decoded daily.
LatestAISoftware & PlatformsCybersecurityStartups & VentureTech Business & Markets
All posts
AIabout 24 hours ago

Inkling 2 Mini brings a looped AI model to local hardware

By megan

Thinking Machines Lab has released open weights for a 26.4-billion-parameter model that runs its middle layers three times per token. Local inference is supported on Apple silicon systems with at least 32GB of memory, but the company warns of limits in factual recall and agent reliability.

A brass cylinder above three circular impressions in clay.

Thinking Machines Lab released Inkling 2 Mini on 20 November 2026 and made its weights available on Hugging Face. The model accepts text and images and produces text. Its distinguishing design choice is to send each token through the same middle layers three times, adding computation without storing three separate copies of those layers.

How the loop works

According to the company’s Inkling 2 Mini model card, the transformer has 24 stored layers. Its first six and last six run once, while the middle 12 run three times. That yields 48 layer applications per token from 24 layers of weights.

This distinction matters when reading the model’s size figures. It has 26.4 billion parameters in total and 3.4 billion active parameters in the underlying network. Reusing the middle layers raises the company’s count to 5.4 billion effective active parameter applications per token. The loop increases work at inference time; it does not add another set of stored weights.

What local use requires

The published model card’s deployment guidance describes a checkpoint of roughly 20GB. Thinking Machines Lab lists an NVIDIA DGX Spark and Apple silicon Macs with at least 32GB of memory as supported local inference options. For other GPUs, it recommends 32GB or more of VRAM. A 20GB checkpoint should therefore not be read as a promise that a machine with only 20GB of memory can run it.

Local inference also needs a compatible runtime and its dependencies. The model card lists Ollama and Exo Labs for Apple silicon, alongside several NVIDIA-oriented frameworks and quantized GGUF options. Developers can alternatively use the company’s Tinker platform. The published requirements describe supported routes, not guaranteed speed or usability on every device.

Where the evidence stops

The release includes company-reported coding benchmark comparisons. Their test setups do not establish how reliably the model will work in a team’s own repository. Thinking Machines Lab also says Mini trails some peers on multi-turn chat and heavily constrained instructions.

Thinking Machines Lab’s risk and limitations section warns that the model can make confident factual errors, repeat unproductive debugging steps, and mishandle detailed project instructions. It may take destructive tool actions without confirmation or report their outcomes inaccurately. For teams evaluating local coding agents, the practical test is their own workload inside a sandbox, with limits on turns and confirmation before irreversible actions.

Continue with this topicAI
Browse categoryarrow_forward
Continue reading

Related posts

Conceptual illustration of a human Mac workspace beside a separate abstract task window with document, app, and web shapes.
AI2 days agoBy megan

Gemini for Mac adds computer use in a limited US rollout

Google says Gemini can work across Mac apps, local files, and websites in a separate background window. The English-language rollout is limited to eligible US personal accounts on paid plans.

Read article
A blank desktop monitor behind a document, clock and stacked cards, with a green permission switch beside them.
AI4 days agoBy megan

Proton launches Lumo Desktop with connectors and scheduled tasks

Lumo now has desktop clients for macOS, Windows and Linux with connectors, schedules and Skills. Proton's setup guide says connector actions default to Auto-approve until changed.

Read article
GeethanTech
GeethanTech
Tech, decoded daily.
Read
Latest postsAISoftware & PlatformsCybersecurityStartups & VentureTech Business & MarketsRSS feed
Connect
mailEmaillanguageWebsiteinLinkedInIGInstagramfFacebook@Xsmart_displayYouTube
From the publisher

Read all our publications: GeethanPost | GeethanTech

A chapter in the ElegantHumanity story

Powered byElegantArc