# Discrete Diffusion Language Models for Interactive Radiology Report Drafting Source: [Discrete Diffusion Language Models for Interactive Radiology Report Drafting]() Collection index: [llms.txt]() Data updated: 2026-07-06T06:00:00.870Z Authors: Max Van Puyvelde, Halil Ibrahim Gulluk, Wim Van Criekinge, Olivier Gevaert Published: 2026-07-01T19:59:09.000Z ## Abstract Diffusion language models, which generate text by denoising a token canvas bidirectionally instead of emitting tokens left to right, have become competitive with autoregressive (AR) generation. Medical foundation models, however, remain almost entirely autoregressive. We adapt a mixture-of-experts diffusion language model, DiffusionGemma-26B, and benchmark it against its same-size AR sibling Gemma-4-26B under an identical LoRA recipe on medical visual question answering datasets, scored by a verbosity-robust LLM judge. Diffusion matches or exceeds AR on all of them, and the finetuned model (3.8B active) is competitive with frontier vision-language models; its decoding is also 3.5-4.4x faster. Beyond this parity, the diffusion model offers a drafting capability AR lacks: any-order infill. Because the canvas is denoised bidirectionally, a radiologist can fix report fragments and have the model fill the text between them, an operation inherent to diffusion but not to autoregression, which is subpar at it. This suits real reports, which are often terse or inconsistent across clinicians and institutions. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Medical Visual Question Answering on SLAKE n=350 (test)](): Medical Visual Question Answering; SLAKE n=350 (test); LLM Judge Accuracy: 86.3. - [Medical Visual Question Answering on VQA-Med 2019 n=350 (test)](): Medical Visual Question Answering; VQA-Med 2019 n=350 (test); Accuracy (LLM Judge): 66.6. - [Medical Visual Question Answering on VQA-RAD n=350 (test)](): Medical Visual Question Answering; VQA-RAD n=350 (test); LLM-Judge Accuracy: 64.9. ## Original sources and implementation [Original paper]() [PDF]()