{"id":"a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371","title":"arXiv cs.AI: WeLM-Image combines image generation, editing and multimodal understanding","argument":"In an arXiv abstract, authors Canmiao Fu and colleagues present WeLM-Image as one model for multimodal understanding, text-to-image generation, instruction-guided editing and multi-reference synthesis. They describe it as designed for single-GPU deployment in WeChat's large-scale consumer applications, processing text, reference images and target latents in one interleaved sequence. Its Mixture-of-Transformers backbone combines shared attention with separate understanding and generation parameters, which the authors say is intended to limit optimization interference. The abstract reports 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token, and describes progressive training up to native 1.5K resolution. The authors report that the 1536-resolution model ranks first among evaluated open-source models on PRISM-Bench, Qwen-Image-Bench and UnicBench. They also describe WeLM-Image-Turbo, a separate version that generates a 1024×1024 image in 4.98 seconds on a single NVIDIA H20 GPU using eight guidance-free steps. These results give the authors’ account of how the unified design performs on the reported comparisons and of a specific speed-and-hardware setup. The abstract does not provide the detailed benchmark results or methods needed to assess those comparisons, or show how they transfer beyond the stated evaluations and setup.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The available evidence is the arXiv abstract; it does not include detailed benchmark results or evaluation methods. The abstract reports a generation-speed result for a specified GPU and step count, not performance across other hardware or deployment settings.","speaker":{"name":"Fu, Canmiao, Li, Chen, Zhang, Guohui, Lu, Hao, Wei, Hongyang, Wu, Huadong, Zhang, Huaisong, Yin, Hubery, Cui, Liyuan, Yan, Qixin, Shao, Tong, Zhou, Xiaoqiang, Leng, Xinhang, Zhang, Ying, Liu, Zhenhai, Huang, Zhipeng","roleAtSource":"Article author"},"source":{"title":"WeLM-Image: A Unified Image Generation Model towards Visual Intelligence","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.07392","publishedAt":"2026-11-05","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors present WeLM-Image as a model integrating multimodal understanding with image generation, editing and multi-reference synthesis.","excerpt":"integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis.","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"},{"text":"The authors say the model is designed for single-GPU deployment in WeChat's large-scale consumer applications.","excerpt":"Designed for single-GPU deployment in WeChat's large-scale consumer applications","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"},{"text":"The abstract reports that the model ranks first among evaluated open-source models on three named benchmarks.","excerpt":"ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench.","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"},{"text":"The authors report WeLM-Image-Turbo's image-generation time and hardware setup.","excerpt":"generates a $1024\\times1024$ image in 4.98\\,s on a single NVIDIA H20 GPU using eight guidance-free steps.","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.07392","checkedAt":"2026-11-08T22:58:07.182Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How do the reported benchmark results compare across the model’s generation and editing tasks?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: WeLM-Image combines image generation, editing and multimodal understanding","summary":"In an arXiv abstract, authors Canmiao Fu and colleagues present WeLM-Image as one model for multimodal understanding, text-to-image generation, instruction-guided editing and multi-reference synthesis. They describe it as designed for single-GPU deployment in WeChat's large-scale consumer applications, processing text, reference images and target latents in one interleaved sequence. Its Mixture-of-Transformers backbone combines shared attention with separate understanding and generation parameters, which the authors say is intended to limit optimization interference. The abstract reports 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token, and describes progressive training up to native 1.5K resolution. The authors report that the 1536-resolution model ranks first among evaluated open-source models on PRISM-Bench, Qwen-Image-Bench and UnicBench. They also describe WeLM-Image-Turbo, a separate version that generates a 1024×1024 image in 4.98 seconds on a single NVIDIA H20 GPU using eight guidance-free steps. These results give the authors’ account of how the unified design performs on the reported comparisons and of a specific speed-and-hardware setup. The abstract does not provide the detailed benchmark results or methods needed to assess those comparisons, or show how they transfer beyond the stated evaluations and setup.","question":"How do the reported benchmark results compare across the model’s generation and editing tasks?","claims":[{"text":"The authors present WeLM-Image as a model integrating multimodal understanding with image generation, editing and multi-reference synthesis.","excerpt":"integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis.","support":"source-stated"},{"text":"The authors say the model is designed for single-GPU deployment in WeChat's large-scale consumer applications.","excerpt":"Designed for single-GPU deployment in WeChat's large-scale consumer applications","support":"source-stated"},{"text":"The abstract reports that the model ranks first among evaluated open-source models on three named benchmarks.","excerpt":"ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench.","support":"source-stated"},{"text":"The authors report WeLM-Image-Turbo's image-generation time and hardware setup.","excerpt":"generates a $1024\\times1024$ image in 4.98\\,s on a single NVIDIA H20 GPU using eight guidance-free steps.","support":"source-stated"}],"limitations":["The available evidence is the arXiv abstract; it does not include detailed benchmark results or evaluation methods.","The abstract reports a generation-speed result for a specified GPU and step count, not performance across other hardware or deployment settings."],"mediaId":"66036ffc304e1c338b942bed3783f542b02afa94cc0476e38c1dda917264b4e0","correctionReason":null,"id":"a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371","publicationId":"13f6f8a3-df45-4eb8-8269-273b8a0dac3b","version":1,"sourceUrl":"https://arxiv.org/abs/2611.07392","sourceTitle":"WeLM-Image: A Unified Image Generation Model towards Visual Intelligence","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-05","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-06T05:07:08.597Z","createdAt":"2026-11-08T21:55:29.851Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"66036ffc304e1c338b942bed3783f542b02afa94cc0476e38c1dda917264b4e0","kind":"article","url":"https://arxiv.org/abs/2611.07392","sourceUrl":"https://arxiv.org/abs/2611.07392","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-08T21:54:36.768Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.07392","checkedAt":"2026-11-08T22:58:07.182Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"66036ffc304e1c338b942bed3783f542b02afa94cc0476e38c1dda917264b4e0","confidence":0.98,"provider":"typesafe","model":"jev-1.13.0","inputHash":"34a9f0ae11f44d78c726cea148065313084d6c1fa7f6c12ed3f68de6b2217f09","decidedAt":"2026-11-08T21:54:37.407Z"},"provenance":{"textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observationId":"6a6394c2cc6e3882fc10b13a26a4b76546a682b2cfa0c767f0d72f87eb2fc61e","retrievedAt":"2026-11-08T21:54:36.667Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"d0fdb770-9e4c-430d-8403-22092470e752"},"reviewer":{"model":"gpt-6-luna","runId":"828a7a23-6130-44bf-881f-df55470dc2b7"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-08T21:55:29.851Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Fu, Canmiao","url":null,"kind":"unknown"},{"name":"Li, Chen","url":null,"kind":"unknown"},{"name":"Zhang, Guohui","url":null,"kind":"unknown"},{"name":"Lu, Hao","url":null,"kind":"unknown"},{"name":"Wei, Hongyang","url":null,"kind":"unknown"},{"name":"Wu, Huadong","url":null,"kind":"unknown"},{"name":"Zhang, Huaisong","url":null,"kind":"unknown"},{"name":"Yin, Hubery","url":null,"kind":"unknown"},{"name":"Cui, Liyuan","url":null,"kind":"unknown"},{"name":"Yan, Qixin","url":null,"kind":"unknown"},{"name":"Shao, Tong","url":null,"kind":"unknown"},{"name":"Zhou, Xiaoqiang","url":null,"kind":"unknown"},{"name":"Leng, Xinhang","url":null,"kind":"unknown"},{"name":"Zhang, Ying","url":null,"kind":"unknown"},{"name":"Liu, Zhenhai","url":null,"kind":"unknown"},{"name":"Huang, Zhipeng","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Fu, Canmiao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Fu, Canmiao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Li, Chen","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Li, Chen","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhang, Guohui","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Guohui","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Lu, Hao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Lu, Hao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Wei, Hongyang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wei, Hongyang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Wu, Huadong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wu, Huadong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhang, Huaisong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Huaisong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Yin, Hubery","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yin, Hubery","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Cui, Liyuan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cui, Liyuan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Yan, Qixin","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yan, Qixin","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Shao, Tong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shao, Tong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:10","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhou, Xiaoqiang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhou, Xiaoqiang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:11","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Leng, Xinhang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Leng, Xinhang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:12","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhang, Ying","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Ying","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:13","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Liu, Zhenhai","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Zhenhai","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:14","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Huang, Zhipeng","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Huang, Zhipeng","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:15","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}}],"participationCoverage":{"total":16,"returned":16,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.07392","snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z","basis":"publisher-page-metadata","sourceLinks":55,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-cs-ai-welm-image-combines-image-generation-editing-and-a99ef496","formats":{"html":"/readings/arxiv-cs-ai-welm-image-combines-image-generation-editing-and-a99ef496","markdown":"/readings/arxiv-cs-ai-welm-image-combines-image-generation-editing-and-a99ef496.md","json":"/readings/arxiv-cs-ai-welm-image-combines-image-generation-editing-and-a99ef496.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-multimodal-model","name":"Multimodal model","explanation":"A model that works with more than one kind of input or output, such as text, images and audio.","referenceUrl":"https://developers.google.com/machine-learning/glossary#multimodal-model"},{"id":"term-gpu","name":"Graphics processing unit","explanation":"A chip that performs many calculations at once; the main hardware used to train and run AI models.","referenceUrl":"https://en.wikipedia.org/wiki/Graphics_processing_unit"},{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"},{"id":"nvidia","name":"NVIDIA","kind":"organization","context":"NVIDIA develops processors, networking and software used for AI, graphics and scientific computing.","referenceUrls":["https://blogs.nvidia.com/"],"path":"/entities/nvidia","json":"/entities/nvidia.json","markdown":"/entities/nvidia.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371","revision":"13f6f8a3-df45-4eb8-8269-273b8a0dac3b","revisionKind":"publication-id","title":"arXiv cs.AI: WeLM-Image combines image generation, editing and multimodal understanding","url":"https://escapemeridian.com/readings/arxiv-cs-ai-welm-image-combines-image-generation-editing-and-a99ef496"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.07392","revision":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","revisionKind":"sha256-source-text","title":"WeLM-Image: A Unified Image Generation Model towards Visual Intelligence","url":"https://arxiv.org/abs/2611.07392"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371","title":"arXiv cs.AI: WeLM-Image combines image generation, editing and multimodal understanding","argument":"In an arXiv abstract, authors Canmiao Fu and colleagues present WeLM-Image as one model for multimodal understanding, text-to-image generation, instruction-guided editing and multi-reference synthesis. They describe it as designed for single-GPU deployment in WeChat's large-scale consumer applications, processing text, reference images and target latents in one interleaved sequence. Its Mixture-of-Transformers backbone combines shared attention with separate understanding and generation parameters, which the authors say is intended to limit optimization interference. The abstract reports 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token, and describes progressive training up to native 1.5K resolution. The authors report that the 1536-resolution model ranks first among evaluated open-source models on PRISM-Bench, Qwen-Image-Bench and UnicBench. They also describe WeLM-Image-Turbo, a separate version that generates a 1024×1024 image in 4.98 seconds on a single NVIDIA H20 GPU using eight guidance-free steps. These results give the authors’ account of how the unified design performs on the reported comparisons and of a specific speed-and-hardware setup. The abstract does not provide the detailed benchmark results or methods needed to assess those comparisons, or show how they transfer beyond the stated evaluations and setup.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The available evidence is the arXiv abstract; it does not include detailed benchmark results or evaluation methods. The abstract reports a generation-speed result for a specified GPU and step count, not performance across other hardware or deployment settings.","speaker":{"name":"Fu, Canmiao, Li, Chen, Zhang, Guohui, Lu, Hao, Wei, Hongyang, Wu, Huadong, Zhang, Huaisong, Yin, Hubery, Cui, Liyuan, Yan, Qixin, Shao, Tong, Zhou, Xiaoqiang, Leng, Xinhang, Zhang, Ying, Liu, Zhenhai, Huang, Zhipeng","roleAtSource":"Article author"},"source":{"title":"WeLM-Image: A Unified Image Generation Model towards Visual Intelligence","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.07392","publishedAt":"2026-11-05","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors present WeLM-Image as a model integrating multimodal understanding with image generation, editing and multi-reference synthesis.","excerpt":"integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis.","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"},{"text":"The authors say the model is designed for single-GPU deployment in WeChat's large-scale consumer applications.","excerpt":"Designed for single-GPU deployment in WeChat's large-scale consumer applications","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"},{"text":"The abstract reports that the model ranks first among evaluated open-source models on three named benchmarks.","excerpt":"ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench.","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"},{"text":"The authors report WeLM-Image-Turbo's image-generation time and hardware setup.","excerpt":"generates a $1024\\times1024$ image in 4.98\\,s on a single NVIDIA H20 GPU using eight guidance-free steps.","url":"https://arxiv.org/abs/2611.07392","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.07392","checkedAt":"2026-11-08T22:58:07.182Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How do the reported benchmark results compare across the model’s generation and editing tasks?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv cs.AI: WeLM-Image combines image generation, editing and multimodal understanding","summary":"In an arXiv abstract, authors Canmiao Fu and colleagues present WeLM-Image as one model for multimodal understanding, text-to-image generation, instruction-guided editing and multi-reference synthesis. They describe it as designed for single-GPU deployment in WeChat's large-scale consumer applications, processing text, reference images and target latents in one interleaved sequence. Its Mixture-of-Transformers backbone combines shared attention with separate understanding and generation parameters, which the authors say is intended to limit optimization interference. The abstract reports 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token, and describes progressive training up to native 1.5K resolution. The authors report that the 1536-resolution model ranks first among evaluated open-source models on PRISM-Bench, Qwen-Image-Bench and UnicBench. They also describe WeLM-Image-Turbo, a separate version that generates a 1024×1024 image in 4.98 seconds on a single NVIDIA H20 GPU using eight guidance-free steps. These results give the authors’ account of how the unified design performs on the reported comparisons and of a specific speed-and-hardware setup. The abstract does not provide the detailed benchmark results or methods needed to assess those comparisons, or show how they transfer beyond the stated evaluations and setup.","question":"How do the reported benchmark results compare across the model’s generation and editing tasks?","claims":[{"text":"The authors present WeLM-Image as a model integrating multimodal understanding with image generation, editing and multi-reference synthesis.","excerpt":"integrates multimodal understanding into text-to-image generation, instruction-guided editing, and multi-reference synthesis.","support":"source-stated"},{"text":"The authors say the model is designed for single-GPU deployment in WeChat's large-scale consumer applications.","excerpt":"Designed for single-GPU deployment in WeChat's large-scale consumer applications","support":"source-stated"},{"text":"The abstract reports that the model ranks first among evaluated open-source models on three named benchmarks.","excerpt":"ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench.","support":"source-stated"},{"text":"The authors report WeLM-Image-Turbo's image-generation time and hardware setup.","excerpt":"generates a $1024\\times1024$ image in 4.98\\,s on a single NVIDIA H20 GPU using eight guidance-free steps.","support":"source-stated"}],"limitations":["The available evidence is the arXiv abstract; it does not include detailed benchmark results or evaluation methods.","The abstract reports a generation-speed result for a specified GPU and step count, not performance across other hardware or deployment settings."],"mediaId":"66036ffc304e1c338b942bed3783f542b02afa94cc0476e38c1dda917264b4e0","correctionReason":null,"id":"a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371","publicationId":"13f6f8a3-df45-4eb8-8269-273b8a0dac3b","version":1,"sourceUrl":"https://arxiv.org/abs/2611.07392","sourceTitle":"WeLM-Image: A Unified Image Generation Model towards Visual Intelligence","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-05","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-06T05:07:08.597Z","createdAt":"2026-11-08T21:55:29.851Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"66036ffc304e1c338b942bed3783f542b02afa94cc0476e38c1dda917264b4e0","kind":"article","url":"https://arxiv.org/abs/2611.07392","sourceUrl":"https://arxiv.org/abs/2611.07392","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-08T21:54:36.768Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.07392","checkedAt":"2026-11-08T22:58:07.182Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"66036ffc304e1c338b942bed3783f542b02afa94cc0476e38c1dda917264b4e0","confidence":0.98,"provider":"typesafe","model":"jev-1.13.0","inputHash":"34a9f0ae11f44d78c726cea148065313084d6c1fa7f6c12ed3f68de6b2217f09","decidedAt":"2026-11-08T21:54:37.407Z"},"provenance":{"textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observationId":"6a6394c2cc6e3882fc10b13a26a4b76546a682b2cfa0c767f0d72f87eb2fc61e","retrievedAt":"2026-11-08T21:54:36.667Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"d0fdb770-9e4c-430d-8403-22092470e752"},"reviewer":{"model":"gpt-6-luna","runId":"828a7a23-6130-44bf-881f-df55470dc2b7"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-08T21:55:29.851Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Fu, Canmiao","url":null,"kind":"unknown"},{"name":"Li, Chen","url":null,"kind":"unknown"},{"name":"Zhang, Guohui","url":null,"kind":"unknown"},{"name":"Lu, Hao","url":null,"kind":"unknown"},{"name":"Wei, Hongyang","url":null,"kind":"unknown"},{"name":"Wu, Huadong","url":null,"kind":"unknown"},{"name":"Zhang, Huaisong","url":null,"kind":"unknown"},{"name":"Yin, Hubery","url":null,"kind":"unknown"},{"name":"Cui, Liyuan","url":null,"kind":"unknown"},{"name":"Yan, Qixin","url":null,"kind":"unknown"},{"name":"Shao, Tong","url":null,"kind":"unknown"},{"name":"Zhou, Xiaoqiang","url":null,"kind":"unknown"},{"name":"Leng, Xinhang","url":null,"kind":"unknown"},{"name":"Zhang, Ying","url":null,"kind":"unknown"},{"name":"Liu, Zhenhai","url":null,"kind":"unknown"},{"name":"Huang, Zhipeng","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Fu, Canmiao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Fu, Canmiao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Li, Chen","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Li, Chen","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhang, Guohui","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Guohui","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Lu, Hao","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Lu, Hao","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Wei, Hongyang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wei, Hongyang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:4","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Wu, Huadong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wu, Huadong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:5","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhang, Huaisong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Huaisong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:6","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Yin, Hubery","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yin, Hubery","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:7","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Cui, Liyuan","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Cui, Liyuan","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:8","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Yan, Qixin","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Yan, Qixin","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:9","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Shao, Tong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Shao, Tong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:10","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhou, Xiaoqiang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhou, Xiaoqiang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:11","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Leng, Xinhang","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Leng, Xinhang","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:12","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Zhang, Ying","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhang, Ying","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:13","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Liu, Zhenhai","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Zhenhai","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:14","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}},{"actor":{"name":"Huang, Zhipeng","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Huang, Zhipeng","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.07392","locator":"head/meta[name=author|citation_author]"},"assertionId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a:participation:15","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-08T22:58:07.182Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","sourceUrl":"https://arxiv.org/abs/2611.07392","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z"}}],"participationCoverage":{"total":16,"returned":16,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.07392","snapshotId":"d31c5b39d516bef18e3156f0b2b3b783ff75504d7d988985d033dbee66ed223a","contentSha256":"f10acb94c0f1290c28d1d08a808574ae4d26c0ce77488d210e12a2199cc3a076","textHash":"0cb3451775a102ee7a8dcc6bd3fb08af8fca5151a47e8c8c585f8b565ad6a95c","observedAt":"2026-11-08T22:58:07.182Z","basis":"publisher-page-metadata","sourceLinks":55,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/a99ef496a1742c54571786f9b3216d60398681c851aac9552c38c6fc0f818371/source"}}}],"audits":[{"publicationId":"13f6f8a3-df45-4eb8-8269-273b8a0dac3b","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading correctly attributes the work to the arXiv paper by Canmiao Fu and colleagues. Design goals, benchmark rankings and speed figures are framed as the authors' reports or descriptions, not as independently established facts.","excerpt":"Authors:Canmiao Fu, Chen Li, Guohui Zhang, Hao Lu"},{"dimension":"summary","status":"uncertain","explanation":"The reading is faithful but largely list-like, and it leaves one central figure hard to interpret. It gives 31.7B stored parameters and about 5.4B activated per target latent token, but omits the abstract's explanation that Mixture-of-Experts modules provide sparse capacity. A non-specialist cannot tell why the two numbers differ. Smallest repair: add that, per the abstract, MoE modules mean only part of the stored parameters are used for each generated token. The reading also drops the abstract's qualifier 'overall' from 'reported overall comparisons'; restoring it would preserve the stated scope.","excerpt":"Equipping each path with Mixture-of-Experts (MoE) modules provides sparse capacity, with 31.7B stored parameters and approximately 5.4B generation parameters activated per target latent token."},{"dimension":"date","status":"pass","explanation":"The originalPublished value of 2026-11-05 matches the arXiv v1 submission date. The feed date of 2026-11-06 is later and is correctly not used.","excerpt":"[Submitted on 5 Nov 2026]"},{"dimension":"support","status":"pass","explanation":"Each claim matches the abstract verbatim in substance, including the scope 'among the evaluated open-source models' and the Turbo speed-and-hardware conditions. The limitations and closing caveat accurately note that the abstract lacks detailed results and methods, without inventing benefits or costs. The question follows from the evidence.","excerpt":"the 1536-resolution model ranks first among the evaluated open-source models on PRISM-Bench, Qwen-Image-Bench, and UnicBench"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified arXiv abstract page, which is the direct original source. It is the most relevant available link.","excerpt":null}],"createdAt":"2026-11-08T21:55:29.851Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}