Distilled Continuous Diffusion Language Models Can Write Code in Few Steps---or One

• URL: http://arxiv.org/abs/2609.04531v1
• Date: Thu, 03 Sep 2026 22:42:57 GMT
• Title: Distilled Continuous Diffusion Language Models Can Write Code in Few Steps---or One
• Abstract summary: We introduce PlaidQ, a 0.7B continuous diffusion language model for code generation.<n>We show that PlaidQ's trajectory can be aggressively distilled into only a few denoising steps.<n>Together, these results establish continuous diffusion as a viable path to few-step and one-step code generation.
• Score: 16.38284188874483
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Language generation is almost universally treated as a sequential process: autoregressive models emit one token at a time, while diffusion language models replace token-level seriality with a long trajectory of iterative refinement. In this work, we introduce PlaidQ, a 0.7B continuous diffusion language model for code generation, and show that its trajectory can be aggressively distilled into only a few denoising steps---or even one, enabling efficient code generation. PlaidQ repurposes a pretrained autoregressive model as a bidirectional denoiser over continuous token embeddings. We distill PlaidQ with distribution matching for few-step generation and paired-trajectory supervision for one-step generation. At matched model scale, PlaidQ is competitive with discrete diffusion language models on code generation. Distillation then shifts the quality--compute frontier: a 16-step student reaches 31.78 and 40.49 pass@10 on HumanEval and MBPP+, surpassing the same PlaidQ teacher sampled for 512 steps. At the extreme, paired-trajectory distillation achieves 7.07 pass@1 on HumanEval with a single denoising step, producing functionally correct programs. Together, these results establish continuous diffusion as a viable path to few-step and one-step code generation. Broadly, continuous diffusion is not merely another representation for language: it provides an interface through which language models can inherit the acceleration and distillation machinery of continuous diffusion modeling. Training and inference code and model checkpoints are available at https://github.com/pengzhangzhi/plaidq.

Related papers

• AURORA-LM: Autoencoding Unified Representation for Continuous-Latent Diffusion Language Modeling [60.04959047453115]
  AURORA-LM is a continuous-latent language model that separates the construction of a decodable text representation from the modeling of its distribution.<n>AURORA-LM achieves the strongest performance among evaluated continuous and diffusion-based language models on OpenWebText free generation and XSum summarization.
  arXiv  Detail & Related papers  (2026-08-03T17:59:50Z)
• DiLaDiff: Distilled Latent-Augmented Diffusion for Language Modeling [51.39835770694148]
  Diffusion language models intrinsically fail to capture correlations between decoded tokens.<n>We propose DiLaDiff, a variant of masked diffusion language models with three components.
  arXiv  Detail & Related papers  (2026-05-22T13:15:59Z)
• Coupling Models for One-Step Discrete Generation [50.37847093567075]
  Coupling Models is a one-step discrete generative model that learns a direct coupling between discrete sequences and Gaussian latents.<n>It avoids complex continuous flows over the simplex and hand-specified data-to-noise couplings.<n>It reduces LM1B text-generation perplexity by 33% at its lowest-perplexity operating point.
  arXiv  Detail & Related papers  (2026-05-08T03:40:39Z)
• LangFlow: Continuous Diffusion Rivals Discrete in Language Modeling [39.014850192441656]
  Continuous diffusion language models (DLMs) lag behind discrete counterparts due to sparse data space and underexplored design space.<n>LangFlow connects embedding-space DLMs to Flow Matching via Bregman divergence.<n>LangFlow rivals top discrete DLMs on both the perplexity (PPL) and the generative perplexity (Gen. PPL)<n>It even exceeds autoregressive baselines in zero-shot transfer on 4 out of 7 benchmarks.
  arXiv  Detail & Related papers  (2026-04-13T17:21:41Z)
• One-step Language Modeling via Continuous Denoising [36.18484491074519]
  We show that language models leveraging flow-based continuous denoising can outperform discrete diffusion in both quality and speed.<n>Our work calls into question the widely held hypothesis that discrete diffusion processes are necessary for generative modeling over discrete modalities.
  arXiv  Detail & Related papers  (2026-02-18T19:23:07Z)
• Transition Matching Distillation for Fast Video Generation [63.1049790376783]
  We present Transition Matching Distillation (TMD), a novel framework for distilling video diffusion models into efficient few-step generators.<n>TMD matches the multi-step denoising trajectory of a diffusion model with a few-step probability transition process.<n>TMD provides a flexible and strong trade-off between generation speed and visual quality.
  arXiv  Detail & Related papers  (2026-01-14T21:30:03Z)
• Continuous Autoregressive Language Models [56.49239051750678]
  We introduce Continuous Autoregressive Language Models (CALM)<n>CALM uses a high-fidelity autoencoder to compress a chunk of K tokens into a single continuous vector.<n>We develop a comprehensive likelihood-free framework that enables robust training, evaluation, and controllable sampling.
  arXiv  Detail & Related papers  (2025-10-31T17:58:11Z)
• FS-DFM: Fast and Accurate Long Text Generation with Few-Step Diffusion Language Models [6.463102424844334]
  We introduce FS-DFM, Few-Step Discrete Flow-Matching.<n>A discrete flow-matching model designed for speed without sacrificing quality.<n>On language modeling benchmarks, FS-DFM with 8 sampling steps achieves perplexity parity with a 1,024-step discrete-flow baseline.
  arXiv  Detail & Related papers  (2025-09-24T23:59:05Z)
• The Diffusion Duality [24.39272541108744]
  Uniform-state diffusion processes naturally emerge from an underlying Gaussian diffusion.<n>Models trained with curriculum learning surpass autoregressive models in zero-shot perplexity on 3 of 7 benchmarks.<n>We present Discrete Consistency Distillation, which adapts consistency distillation from the continuous to the discrete setting.
  arXiv  Detail & Related papers  (2025-06-12T16:55:35Z)
• FlashDLM: Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusion [22.207275433870937]
  Diffusion language models offer parallel token generation and inherent bidirectionality.<n>State-of-the-art diffusion models (e.g., Dream 7B, LLaDA 8B) suffer from slow inference.<n>We introduce Guided Diffusion, a training-free method that uses a lightweight pretrained autoregressive model to supervise token unmasking.
  arXiv  Detail & Related papers  (2025-05-27T17:39:39Z)
• DiffusionBERT: Improving Generative Masked Language Models with Diffusion Models [81.84866217721361]
  DiffusionBERT is a new generative masked language model based on discrete diffusion models. We propose a new noise schedule for the forward diffusion process that controls the degree of noise added at each step. Experiments on unconditional text generation demonstrate that DiffusionBERT achieves significant improvement over existing diffusion models for text.
  arXiv  Detail & Related papers  (2022-11-28T03:25:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.