Intel / article

ALoDLM 1.7B released

Source: huggingface.co
Author: Amazon model releases
Date: Published Oct 2, 2026

Context

Amazon's ALoDLM-1.7B is the smaller sibling of ALoDLM-8B, a diffusion : instead of writing strictly left to right, it settles several token positions together. The model card says it is initialized from Qwen3-1.7B (from Qwen). Confident tokens are committed early and serve as , while unresolved tokens keep their hidden states and pass through the shared layers again until a learned halting policy stops the loop.

Amazon reports an unweighted average over eleven benchmarks of 65.5, against 63.8 for Qwen3-1.7B and 61.0 for SDAR-1.7B, a third model the card compares against. That is a gain of under two points over the base model, on the company's own evaluation. The card reports throughput for the 8B model only, so the 1.7B model's speed is not established by this source. The license is noncommercial.

Terms in this piece · Glossary

LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
transformer — The neural network architecture behind modern AI models, built on attention — letting every word directly consider every other word in parallel.
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.

Read the source huggingface.co