AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

MoSs: Mixture of Scales for Efficient High-Resolution Autoregressive Image Generation

Yaoxiu Lian, Hao Liang, Zhihong Gou, Yijia Zhang, Jiaming Xu, Guohao Dai, Ningyi Xu

AAAI 2026 · Image generation › Autoregressive and latent generation · approach: Next-scale and token autoregressive images · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
apply MoSs5 parallel refinementInfinity-2B · Infinity-2B 1024x1024latency (s): 1.29 -> 0.26 (4.96x speedup)Improved
apply MoSs5Infinity-2B · Infinity-2BKV cache (GB): 3.13 -> 0.45Improved
apply MoSs5 (more parallel scales)Infinity-2B · GenEvaloverall: 0.729 -> 0.705Hurt
apply MoSs2Infinity-2B · GenEvaloverall: 0.729 -> 0.728No clear effect
apply MoSs2HART-0.7B · 1024x1024speedup: 1.56x speedup; KV cache 2.79 -> 0.90 GBImproved
apply MoSs2HART-0.7B · GenEvaloverall: 0.498 -> 0.488Hurt
add Cross-Scale Self-Attention instead of standard self-attention (ablation)MoSs5 · GenEvaloverall: 0.674 -> 0.705Improved
add Cross-Scale Self-Attention instead of standard self-attention (ablation)MoSs2 · GenEvaloverall: 0.720 -> 0.728Improved
add Triton RoPE2D kernel (ablation)MoSs with CSSA · Infinity-2Bspeedup: 1.87 -> 2.05 (MoSs2)Improved
MoSs3 vs FastVARInfinity-2B · Infinity-2BGenEval overall / speedup: 0.724 vs 0.721; both 2.87xNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy