Loading
Life sciences · Preprint
arXiv · September 3, 2026
Early or partial results. Treat as a signal, not a conclusion.
This preprint describes Uno, a diffusion-augmented LLM architecture combining autoregressive and discrete diffusion components to accelerate parallel token generation. The authors report up to 3× speedups over baseline autoregressive models and claim benchmark performance competitive with or superior to existing acceleration and diffusion-based methods, but the work lacks peer review and independent validation.
Empirical method development and benchmarking study. Benchmark evaluation on agentic tool use, coding, and long-context reasoning tasks; no human participants or clinical population.. Intervention: Diffusion-augmented LLMs (Uno) with Ψ-Spec samplers for parallel token generation, decoupling autoregressive and diffusion-trained weights. Compared with: Speculative decoding methods, diffusion LLMs (DiffusionGemma), proprietary Mercury 2, and base autoregressive models.
Uno achieves up to 3× speedups over base AR model across all evaluated batch sizes 8B Uno model outperforms 26B DiffusionGemma and proprietary Mercury 2 on agentic tool use, coding, and long-context reasoning benchmarks Higher throughput than leading speculative-decoding methods at every evaluated batch size
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
A novel method for accelerating LLM inference with reported speedups and benchmark comparisons, but presented as a preprint without peer review, limiting strength of evidence for clinical or production deployment claims.
As stated by the source record.
Quoted from the source exactly as published.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
Large Language Models (LLMs) owe much of their success to next-token prediction (NTP), but their autoregressive (AR) structure requires slow, sequential token generation. To overcome this bottleneck, we introduce diffusion-augmented LLMs, a new class of models that defines an AR model distribution while using diffusion to draw multiple tokens in parallel from that distribution. We decouple the parameters of these models into two sets: AR weights, trained using the standard NTP objective, and lightweight diffusion weights, trained to generate multiple tokens simultaneously. The diffusion weights are learned through a simple Diffusion Distillation phase that adds negligible overhead to existing LLM training pipelines. We also introduce $Ψ$-Spec, a family of samplers that enables lossless acceleration and inference-time scaling at a fixed context length. Unlike speculative decoding, our method requires no separate draft model. Unlike diffusion LLMs (d-LLMs), it accelerates generation without sacrificing the quality of the underlying AR model. The resulting models, called Uno, can be trained from scratch or built by augmenting existing open-weight AR LLMs. Uno achieves higher throughput than leading speculative-decoding methods at every evaluated batch size and delivers up to $3\times$ speedups over the base AR model, including at the largest batch size supported by the device. Notably, our 8B Uno model outperforms the leading open d-LLM, the 26B DiffusionGemma, and the proprietary Mercury 2 across all evaluated benchmarks in agentic tool use, coding, and long-context reasoning. We release code and checkpoints at: https://s-sahoo.github.io/uno/
Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.