CanvasFlash: Efficient Autoregressive Image Generation with Semantic Representation

Long visual sequences impose substantial training and decoding costs on high-resolution autoregressive (AR) image generation. We propose CanvasFlash, a novel approach that reduces this cost by exploiting the redundancy within visual representations. We depart from the reconstruction-oriented design in favor of a semantic image tokenizer that serves as a high-level information bottleneck, and delegate pixel rendering to a diffusion decoder. The resulting token distribution becomes highly concentrated: diverse tokens cluster in semantically informative regions, whereas broad regions share a common Canvas Token, indicating that only a fraction of the sequence is needed to characterize an image for semantic representation. Leveraging this sparsity, we propose Canvas-Sparse Training, which processes $41\%$ of the tokens required by full-sequence training. We also propose Canvas-as-Draft Decoding, which uses the Canvas Token as a parameter-free draft and achieves up to $3.09\times$ speedup in AR decoding. With reinforcement learning, CanvasFlash combines these efficiency gains with competitive image generation performance. More broadly, CanvasFlash highlights the redundancy carried by reconstruction-oriented visual representations, and validates semantic tokenization as a practical route toward efficient autoregressive image generation.