Friday, 13 November 2026, story 6 of 7
Granite 4.1 3B and 8B models get LLMON, an LLM-native markup language
A Hugging Face post says the markup is now in both models, but the captured page gives no syntax, license or usage details.

What happened
A Hugging Face post says LLMON, described in its title as an LLM-native markup language, is now in the Granite 4.1 3B and 8B models. The captured page holds no article body, only navigation and listings.
It shows three linked model entries labelled 4B, 8B and 9B. Two were updated about 11 hours ago and one on Apr 28. It also links a Space called LLMON Attention Visualization, which invites users to explore live Granite LLMON instruction attention. The page does not give a license, a price or a description of the format.
What it means
The page confirms a name and two model sizes, not how LLMON works. It does not say whether the models were trained on the format, whether it is a prompt convention, or whether existing prompts keep working. Those answers decide whether this matters to anyone running small models locally.
The linked attention Space hints at the intent. If LLMON is meant to mark which parts of an input are instructions, the authors may be offering a way to inspect how a model weighs them. That is an inference from a Space title, not a stated claim. The mismatch between the 3B in the title and the 4B and 9B labels on the listings is also unexplained. Builders should read the full post and model cards before testing.
What to do: Wait for the full post and model cards before changing prompts or pipelines; the available text confirms the name and the models, not how LLMON works.
Source
Filed under Open source. Published .
"What it means" is analysis, not reporting. Found an error? Write to daily@vibitely.com.