{"paper":{"title":"Hunyuan3D 2.1: From Images to High-Fidelity 3D Assets with Production-Ready PBR Material","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Hunyuan3D 2.1 generates high-fidelity 3D assets with production-ready PBR materials from images using two dedicated models.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bojian Zheng, Bowen Zhang, Chao Zhang, Chunchao Guo, Di Luo, Di Wang, Dongyuan Guo, Haolin Liu, Hao Zhang, Hongyu Yan, Huiwen Shi, Jiaao Yu, Jianchen Zhu, Jie Jiang, Jihong Zhang, Jingwei Huang, Junlin Yu, Kai Liu, Liang Dong, Lifu Wang, Lin Niu, Linus, Meng Chen, Mingxin Yang, Peng Chen, Peng He, Qingxiang Lin, Runzhou Wu, Sheng Zhang, Shida Wei, Shilin Chen, Shirui Huang, Shuhui Yang, Shu Liu, Sicong Liu, Team Hunyuan3D, Tian Liu, Xianghui Yang, Xiang Yuan, Xiaofeng Yang, Xin Huang, Yifei Feng, Yifu Sun, Yiwen Jia, Yixuan Tang, Yonghao Tan, Yuhong Liu, Yulin Cai, Yunfei Zhao, Zebin He, Zeqiang Lai, Zheng Ye, Zibo Zhao","submitted_at":"2025-06-18T13:14:46Z","abstract_excerpt":"3D AI-generated content (AIGC) is a passionate field that has significantly accelerated the creation of 3D models in gaming, film, and design. Despite the development of several groundbreaking models that have revolutionized 3D generation, the field remains largely accessible only to researchers, developers, and designers due to the complexities involved in collecting, processing, and training 3D models. To address these challenges, we introduce Hunyuan3D 2.1 as a case study in this tutorial. This tutorial offers a comprehensive, step-by-step guide on processing 3D data, training a 3D generati"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The system comprises two core components: the Hunyuan3D-DiT for shape generation and the Hunyuan3D-Paint for texture synthesis.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the outlined data preparation, architecture, and training strategies will reliably produce high-resolution, production-ready 3D assets with PBR materials.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Hunyuan3D 2.1 is a two-part system with DiT for shape generation and Paint for texture synthesis that produces high-fidelity 3D assets with PBR materials.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Hunyuan3D 2.1 generates high-fidelity 3D assets with production-ready PBR materials from images using two dedicated models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"492723220de3d623e5ca8ed616242a88b04fe05f458d4c2acc842df93cf7ded5"},"source":{"id":"2506.15442","kind":"arxiv","version":1},"verdict":{"id":"11fa80ec-7012-48a4-9453-7d7e2cc0651d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T23:05:38.147652Z","strongest_claim":"The system comprises two core components: the Hunyuan3D-DiT for shape generation and the Hunyuan3D-Paint for texture synthesis.","one_line_summary":"Hunyuan3D 2.1 is a two-part system with DiT for shape generation and Paint for texture synthesis that produces high-fidelity 3D assets with PBR materials.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the outlined data preparation, architecture, and training strategies will reliably produce high-resolution, production-ready 3D assets with PBR materials.","pith_extraction_headline":"Hunyuan3D 2.1 generates high-fidelity 3D assets with production-ready PBR materials from images using two dedicated models."},"references":{"count":47,"sample":[{"doi":"","year":2020,"title":"Denoising diffusion probabilistic models","work_id":"80d27c68-54aa-49bf-b4e5-f94cb73ea612","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"High-resolution image synthesis with latent diffusion models","work_id":"4055c382-1505-417a-a4bf-813919c20dc1","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Scaling rectified flow transformers for high-resolution image synthesis","work_id":"67cbf2e9-fbed-420e-b994-f62b956bd45e","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Hunyuan-dit: A powerful multi-resolution diffusion transformer with fine-grained chinese understanding","work_id":"b838659f-1306-4224-b90d-4a4f84b2369c","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Hunyuanvideo: A systematic framework for large video generative models, 2024","work_id":"ab864055-7e76-41c8-98d8-bedfb22e7384","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":47,"snapshot_sha256":"82d25d581316d9695bf1985d7d0bf5e66d226cf3028400bb7437c54217687e8a","internal_anchors":13},"formal_canon":{"evidence_count":2,"snapshot_sha256":"b6ac97025744070b569156ff0f86a55cb65e1170985bda7408890f0861f4a40e"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}