TOPIC · VISUAL-GEN · 67 PAPERS
LLMsReasoningAlignmentAgentsMultimodalVisual generationVision & 3DRobotics & embodied AIEfficient training & inferenceSafety & interpretabilityAI for ScienceCode & SWEBenchmarks & dataReinforcement learning
#03 ▲223 ★12k ❝1 2026-09-27
YuE2 plans a readable score first and then renders full songs, unifying symbolic and audio music generation.
Multimodal Art Projection · Ruibin Yuan, Jiahao Pan, Junyan Jiang +32
#16 ▲161 ★512 ❝1 2026-09-21
WorldCrafter adds a camera-queryable implicit 3D memory to a video world model to keep long explorations consistent.
ARC Lab, Tencent · Wangbo Yu, Kunhao Liu, Wenbo Hu +8
#18 ▲364 ★312 ❝0 2026-10-14
ATAS compiles screenplays into shots backed by scene and event memories so long agentic videos stay spatially and temporally consistent.
Zhejiang University · Weijie Wang, Haomin Zhang, Ye Li +7
#20 ▲293 ★201 ❝1 2026-10-08
AgentGarten couples simulators and game engines with a neural renderer to build faithful, realistic worlds for training agents.
MirroS · Jiawei Chi, Shangchen Miao, Zhiyuan Shi +12
#24 ▲325 ★154 ❝0 2026-10-08
DiffusionART aggregates rewards before tilting the objective, making online RL fine-tuning of diffusion models efficient and stable.
KlingAI Research · Zhenglin Wan, Jiayu Chen, Zhengfeiyang Zhang +6
#28 ▲367 ★36 ❝1 2026-09-28
MaLiang-Harness has MLLMs generate images and videos via programs in a build-inspect-revise loop that closes the program-to-visual gap.
National University of Singapore · Haoyu Zhao, Zihao Zhang, Xudong Wang +4
#39 ▲174 ★364 ❝0 2026-10-04
Kandinsky 6.0 Video generates 5-second clips with synchronised 44 kHz audio and lip-sync, in 3B and 29B variants.
Kandinsky Lab · Team Kandinsky, Julia Agafonova, Bulat Akhmatov +85
#40 ▲167 ★97 ❝0 2026-09-23
Builds 150 cognitive-science tasks to test and train object permanence in video world models.
Carnegie Mellon University · Haotian Zhang, Fengyuan Yu, Dezhi Luo +28
#44 ▲197 ★123 ❝0 2026-10-07
UltraText Bench uses 432 bilingual prompts to evaluate dense visual text rendering in image generators.
Westlake University · Deyuan Liu, Yihao Hu, Jingxuan Zhang +13
#46 ▲137 ★114 ❝0 2026-10-13
AesCode uses a generated image as aesthetic guidance and decoupled cross-modal rewards to produce better-looking slides and posters as code.
Microsoft · Xin Gao, Yifan Yang, Xin Li +13
Code & SWEVisual generationReinforcement learning arXiv Code HF
#49 ▲131 ★523 ❝0 2026-10-14
InfiniSplat v2 analyses and improves feed-forward 3D Gaussian reconstruction for better rendering under large viewpoint changes.
zju3dv · Yongzhen Hu, Hao Yu, Jiawei Wang +7
#50 ▲78 ★526 ❝0 2026-09-21
GAE reparameterises geometry foundation-model features into a compact latent space so generated worlds stay 3D-consistent.
ARC Lab, Tencent · Jiahao Lu, Minghao Yin, Wenbo Hu +5
#56 ▲89 ★1.2k ❝0 2026-10-16
MOSS-Image is an open 5B text-to-image system strong at bilingual long-text rendering and layouts, with intent-aware prompt enhancement.
OpenMOSS · MOSI. AI
#57 ▲101 ★56 ❝1 2026-10-02
DuoMatching matches both joint and per-frame marginal distributions to improve quality and alignment in few-step video generation.
ByteDance · Jiahao Zhan, Yan Wang, Yongrui Ma +5
Visual generationEfficient training & inference arXiv Code HF
#72 ▲179 ★36 ❝0 2026-10-19
Ropedia W0 is an egocentric loco-manipulation world foundation model spanning humans and robots, trained on 10k hours of human and 3k hours of robot data.
Ropedia · Tianqi Liu, Hanfei Zhao, Qifeng Liang +12
#74 ▲117 ★185 ❝0 2026-10-15
3DCodeVerse offers 2M executable 3D programs and a coding harness so agents can create editable, verifiable 3D content.
University of Southern California · Yipeng Gao, Ziyao Zeng, Jinfa Huang +16
#76 ▲63 ★210 ❝3 2026-09-30
A survey of post-training and alignment for video generation, covering temporal coherence, physics and safety challenges.
Arizona State University · Chaoyu Li, Xiaoyi Gu, Yogesh Kulkarni +10
#85 ▲138 ★17 ❝1 2026-09-25
FuseReg learns regularised layer fusion to ease the reconstruction-generation trade-off in representation autoencoders.
USC Physical Superintelligence (PSI) Lab · Hongyang Du, Yunfei Xie, Junjie Ye +13
#94 ▲74 ★100 ❝0 2026-10-12
Multics generates multi-player world videos with hierarchical player-group-global communication that scales to more players.
ARC Lab, Tencent · Kaiyi Huang, Wangbo Yu, Yukun Huang +6
#97 ▲74 ★49 ❝3 2026-10-01
PROWBench tests whether video models faithfully render program-specified world events, aimed at next-gen game engines.
Alaya Lab · Zheng-Hui Huang, Guixu Lin, Yu-Ju Tsai +6
#106 ▲124 ★30 ❝0 2026-09-30
FrameMorrow selects history frames by future relevance via prospective tokens for efficient, consistent long video generation.
National University of Singapore · Bo Yin, Xiaobin Hu, Jiaqi Zhao +1
Visual generationEfficient training & inference arXiv Code HF
#114 ▲102 ★12 ❝0 2026-10-07
VIS-Ground lets viewers steer interactive video stories while staying consistent with the source and rendered history.
Google · Bingxuan Li, Yiwen Song, Xueqing Wu +10
#124 ▲104 ★35 ❝0 2026-09-29
The Thinking Reward Model first sets case-specific rubrics and then scores, giving finer rewards for visual generation.
Affiliation unverified · Xuehai Bai, Zhenchen Tang, Yang Shi +9
#126 ▲52 ★103 ❝0 2026-10-13
LensErase is a general erase-anything model trained on 100M pairs, removing objects, text, watermarks and their visual effects.
Microsoft · Yunuo Chen, Xiuyu Wu, Qinhong Yang +11
#127 ▲34 ★227 ❝1 2026-09-29
PixelUMM unifies image and video understanding and generation directly in pixel space without a visual encoder.
NVIDIA · Cong Wei, Xuanchi Ren, Bryan Chu +8
#135 ▲47 ★215 ❝1 2026-10-20
SolarRenderer adds two LoRAs to a frozen video model to render coarse structural video into realistic footage in four steps, including real-time streaming.
Show Lab · Guian Fang, Shengju Qian, Dongyue Lu +6
Visual generationEfficient training & inference arXiv Code HF
#136 ▲60 ★207 ❝0 2026-10-07
SGF+ separates parameters for context writing and denoising in autoregressive video generation to resolve conflicting gradients.
Affiliation unverified · Zihan Su, Junhao Zhuang, Yaowei Li +10
#140 ▲89 ★34 ❝0 2026-10-01
World Observer jointly generates panoramic observer views alongside the actor view so off-screen objects keep evolving consistently.
KAIST AI · Hyunwook Choi, Dahyun Chung, Hyunsung Kim +4
#143 ▲51 ★81 ❝0 2026-10-14
Intern Lumina U2 is a multi-codebook diffusion LLM unifying visual understanding and image generation in one space.
Shanghai AI Laboratory · Qian Yu, Zongyang Qiu, Juncheng Yan +40
#145 ▲43 ★48 ❝1 2026-10-14
Orbia is a benchmark for 3D consistency of interactive world models, testing whether scenes persist on revisits.
Tencent · Dongyue Lu, Guian Fang, Hur Lim +20
#149 ▲64 ★44 ❝1 2026-10-08
OuroWorld turns any static 3D Gaussian scene into a seamlessly looping 3D cinemagraph viewable from any angle.
Alaya Lab · You-Zhe Xie, Ting-Wei Chou, Yu-Hsuan Li +3
#150 ▲140 ❝0 2026-09-29
Adaptive Reward Routing dynamically adjusts where updates act and how rewards combine to improve joint audio-video diffusion.
Tencent · Songlin Yang, Xiaotong Zhao, Jiacheng Zhang +5
#153 ▲90 ★26 ❝0 2026-10-07
GRACE further compresses the video latent while staying compatible with the pretrained DiT, speeding up video generation.
KAIST AI · Jiyoung Kim, Paul Hyunbin Cho, Jisu Nam +5
Visual generationEfficient training & inference arXiv Code HF
#159 ▲93 ★21 ❝0 2026-10-15
Speculative Forcing treats intermediate estimates as drafts and accepts converged tokens early to speed up autoregressive video generation.
NAVER AI Lab · Hyunwook Choi, Seungryong Kim, Dongyoon Han +1
Visual generationEfficient training & inference arXiv Code HF
#162 ▲72 ★21 ❝1 2026-10-08
A multi-agent egocentric world model that generates synchronised first-person streams of agents interacting in a shared world.
KAIST AI · Dahyun Chung, Siyoon Jin, Hyunwook Choi +5
#164 ▲66 ★89 ❝0 2026-09-23
Examines memory for autoregressive video generation: keeping identities, states and causal changes within bounded context.
Affiliation unverified · Harold Haodong Chen, Rongjin Guo, Disen Lan +22
#166 ▲21 ★619 ❝1 2026-10-04
Prism is dynamic sparse attention that focuses on sound-producing regions for training native 2K joint video-audio models.
Tencent Hunyuan · Shuyuan Tu, Qi Tian, Yinming Huang +8
Visual generationEfficient training & inference arXiv Code HF
#168 ▲38 ★625 ❝0 2026-09-29
HelixWorld is a real-time interactive world model where visuals and spatial stereo sound co-evolve with user input.
NoizAI · Lei Ke, Jiahao Pan, Zeyue Tian +13
#179 ▲46 ★46 ❝0 2026-10-05
MC-Sparse is training-free sparse attention that selects individual KV tokens with cached compensation, closing the dense-sparse gap in DiTs.
Tencent Hunyuan · Jiarui Chen, Zeqiang Lai, Jiangshan Wang +5
Visual generationEfficient training & inference arXiv Code HF
#181 ▲182 ❝0 2026-10-20
Matrix-Game 4.0 separates game logic (engine plus coding agent) from visuals (autoregressive video renderer), reaching playable 60 FPS on one RTX 5090.
Affiliation unverified · Jihai Zhang, Runjia Qian, Kaiqin Kong +15
#188 ▲63 ★11 ❝0 2026-10-01
AutoGUIWorld uses image generators as visual world models of GUIs to synthesise diverse interaction trajectories.
Tencent Hunyuan · Cheng Yang, Yifan Wu, Yutao Huang +18
#195 ▲41 ★233 ❝0 2026-10-06
WorldSonus adds real-time, spatially aligned stereo sound to world models with mid-stream sound control.
NoizAI · Pengjun Fang, Jingyi Fa, Kam Man Wu +9
#199 ▲84 ★10 ❝0 2026-10-14
Speculative Sampling for interactive video generation caches predictions from large windows to get large-chunk quality at small-chunk latency.
university of science and technology of china · Yubo Huang, Yuxuan Zhang, Jiaming Liu +2
Visual generationEfficient training & inference arXiv Code HF
#203 ▲37 ★67 ❝0 2026-10-09
Trains a corrector that refines the denoiser's prediction at each step to reduce exposure bias, improving DiT, SiT and JiT samples.
University of California, Berkeley · Donghoon Ahn, Runqian Wang, Alexei A. Efros +2
#213 ▲104 ❝1 2026-09-21
RULER uses instance-aware multi-axis rubrics as rewards for SVG generation, reducing the reward hacking seen with scalar metrics.
inclusionAI · Hangyu Ran, Yuhao Zheng, Yingying Zhang +2
#217 ▲58 ★38 ❝0 2026-10-12
AReA adapts the attention receptive field across denoising steps so diffusion transformers generate beyond-training resolutions more cheaply.
University of Toronto Computer Science · Javad Rajabi, Koorosh Roohi, David B. Lindell +1
Visual generationEfficient training & inference arXiv Code HF
#225 ▲107 ❝1 2026-10-01
SALD lets a diffusion model score each image-caption pair at two noise levels, yielding self-referenced training signals without a teacher.
Indian Institute of Technology, Kanpur · Aryan Das, Surjo Dey, Koushik Biswas +4
#227 ▲68 ★14 ❝0 2026-09-30
StereoBind ties stereo sound position to on-screen source motion for immersive joint video-audio generation.
Affiliation unverified · Hanmo Chen, Chengcheng Liu, Tianxiao Chen +7
#231 ▲41 ★123 ❝0 2026-10-13
FlashVSR+ speeds up one-step video super-resolution with hybrid attention and aligns it to human preference via reward-weighted distribution matching.
MMLab-CUHK · Ruchang Yao, Shi Guo, Junhao Zhuang +7
Visual generationEfficient training & inference arXiv Code HF
#244 ▲86 ★4 ❝0 2026-10-15
SCOPE, a GRPO-based objective, credits semantic spans rather than sub-word tokens when training prompt rewriters for text-to-image.
University of Rochester · Yufei Jiao, Jinfa Huang, Zeyue Xue +5
#251 ▲60 ★24 ❝0 2026-10-01
Uses MLLM spatial understanding to manage long-term memory so world models stay spatially consistent in long video generation.
Affiliation unverified · Ying Yang, Guiyu Zhang, Lianghua Huang +5
#253 ▲50 ★31 ❝0 2026-09-28
Interleaved RL jointly trains reflection text and image generation so unified models can diagnose and repair their own images.
MMLab@NTU · Yijia Fan, Ziqi Huang, Zhongang Cai +5
MultimodalReinforcement learningVisual generation arXiv Code HF
#258 ▲39 ★28 ❝1 2026-10-01
Stability AI's 4Director controls camera and object motion in a video world model via explicit rigid meshes and per-frame transforms.
Stability AI · Wei Cao, Hao Zhang, Vikram Voleti +5
#261 ▲31 ★194 ❝0 2026-10-07
Iris-3B pretrains and converts pixel-space diffusion models to test whether skipping the VAE helps detail-sensitive tasks.
speridlabs · Hanqiu Li Cai, Chema Garabito
#265 ▲33 ★16 ❝0 2026-10-09
Studies which visual representations actually help pixel diffusion transformers, finding better understanding does not imply better generation.
alibaba · Zihao Pan, Mengting Chen, Cheng Shi +6
#268 ▲73 ❝0 2026-09-21
VideoGen-Agent trains a multimodal agent with multitask agentic RL to use tools for knowledge- and identity-heavy video prompts.
Princeton University · Binxu Li, Haoyi Duan, Yuhui Zhang +9
#280 ▲42 ★17 ❝0 2026-10-20
VPM is a world model that jointly forecasts multi-view future video and 3D point trajectories, extended to robot control.
Affiliation unverified · Byungwoo Jeon, Sunghwan Hong, Dohyeon Kim +10
#285 ▲37 ★22 2026-10-18
MT Forcing trains autoregressive-diffusion editors with a single-turn distillation teacher and teacher-forcing guidance to curb multi-turn editing drift.
SenseNova · Houyuan Chen, Fanzhou Wang, Xiangyu Fan +10
#286 ▲27 ★8 2026-10-19
RoAR-VAE trains tokenizers so semantics and pixels remain recoverable from corrupted latents, making latents easier for diffusion models to learn.
Tsinghua-LeapLab · Yang Yue, Fangyun Wei, Yitong Wang +4
#287 ▲26 ★47 ❝0 2026-10-19
Manga2Animate uses paired creator and reviewer agents to plan scripts and storyboards before generating animation faithful to the source manga.
National Yang Ming Chiao Tung University · Sheng-Wei Zhang, Hsi-An Chen, Yan-Ting Liou +3
#288 ▲33 ★12 ❝0 2026-10-16
MIST separates instruction execution from memory in video world model tests, finding the bottleneck is usually execution rather than forgetting.
The University of Hong Kong · Weiyang Jin, Jiwen Yu, Sihan Xu +2
#295 ▲27 ❝0 2026-10-19
SWIFT extrapolates training targets away from a frozen weak denoiser to boost a stronger diffusion model with no extra inference cost.
alibaba · Yanjin He, Cuifeng Shen, Wanzhou Lei +5
#297 ▲22 ★9 ❝0 2026-10-06
A data-free, simulation-free diffusion distillation where one student both generates and estimates scores, reaching strong one-step FID.
Vector Institute · Yuxiang Fu, Qi Yan, Zike Wu +4
Visual generationEfficient training & inference arXiv Code HF
#302 ▲22 ❝0 2026-10-20
Co-generates video, replayable 3D motion for any rig and a registering camera, turning video-model motion into usable 3D animation.
NVIDIA · Noam Rotstein, Yoad Tewel, Chen Tessler +5
#303 ▲33 ❝0 2026-10-16
MTEdit builds 500K multi-turn image-editing trajectories by generating several edit states in one packed canvas, plus a benchmark.
SenseNova · Houyuan Chen, Fanzhou Wang, Xiangyu Fan +8
#305 ▲26 ❝0 2026-10-03
Argues real-world super-resolution works best in a frozen vision-foundation feature space rather than reconstruction latents or pixels.
RTP-LLM · Wanzhou Lei, Cuifeng Shen, Yanjin He +6
#307 ▲22 2026-10-19
SpecKCast forecasts DiT outputs in the frequency domain with Kalman filters, running the model only when uncertainty is high, for training-free speedup.
Samsung Research · Javad Rajabi, Hakki Can Karaimer, David B. Lindell +2
Data updated: Latest AI papers(UTC+8)