{"paper":{"title":"Visual ChatGPT: Talking, Drawing and Editing with Visual Foundation Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Visual ChatGPT lets users chat with images by linking ChatGPT to visual foundation models through prompts.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenfei Wu, Nan Duan, Shengming Yin, Weizhen Qi, Xiaodong Wang, Zecheng Tang","submitted_at":"2023-03-08T15:50:02Z","abstract_excerpt":"ChatGPT is attracting a cross-field interest as it provides a language interface with remarkable conversational competency and reasoning capabilities across many domains. However, since ChatGPT is trained with languages, it is currently not capable of processing or generating images from the visual world. At the same time, Visual Foundation Models, such as Visual Transformers or Stable Diffusion, although showing great visual understanding and generation capabilities, they are only experts on specific tasks with one-round fixed inputs and outputs. To this end, We build a system called \\textbf{"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We build a system called Visual ChatGPT, incorporating different Visual Foundation Models, to enable the user to interact with ChatGPT by 1) sending and receiving not only languages but also images 2) providing complex visual questions or visual editing instructions that require the collaboration of multiple AI models with multi-steps.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That prompt-based injection of visual model capabilities into ChatGPT enables reliable multi-step collaboration without frequent errors in task decomposition or model selection.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Visual ChatGPT integrates visual foundation models with ChatGPT via prompts to enable multi-step image understanding, generation, and editing in conversational interactions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Visual ChatGPT lets users chat with images by linking ChatGPT to visual foundation models through prompts.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"5f333e5e5df926a0260980b6126e0f036c3ff09b2748a6e0f927dba23873a111"},"source":{"id":"2303.04671","kind":"arxiv","version":1},"verdict":{"id":"be73130e-f746-4664-9b0e-2a5ad3d993da","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T22:46:45.409554Z","strongest_claim":"We build a system called Visual ChatGPT, incorporating different Visual Foundation Models, to enable the user to interact with ChatGPT by 1) sending and receiving not only languages but also images 2) providing complex visual questions or visual editing instructions that require the collaboration of multiple AI models with multi-steps.","one_line_summary":"Visual ChatGPT integrates visual foundation models with ChatGPT via prompts to enable multi-step image understanding, generation, and editing in conversational interactions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That prompt-based injection of visual model capabilities into ChatGPT enables reliable multi-step collaboration without frequent errors in task decomposition or model selection.","pith_extraction_headline":"Visual ChatGPT lets users chat with images by linking ChatGPT to visual foundation models through prompts."},"references":{"count":58,"sample":[{"doi":"","year":2022,"title":"Flamingo: a visual language model for few-shot learning","work_id":"059e2edb-7251-4c10-907e-c021375c785d","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2015,"title":"Vqa: Visual question answering","work_id":"cbdd1a33-3045-4f80-9a99-f0866c9fa2ac","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Vlmo: Unified vision-language pre- training with mixture-of-modality-experts","work_id":"bbdebaac-4da3-44ee-92f1-4125efdfe5d8","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"In- structpix2pix: Learning to follow image editing instructions","work_id":"de9c4a95-36ff-4e97-a0c0-3cca39f60b1a","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1901,"title":"Lan- guage models are few-shot learners","work_id":"586b225f-281e-4de8-8c7c-dd02e7091ca8","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":58,"snapshot_sha256":"0733ffc7970d1b55320fd6d78169072f7d61c0e58b284ff49a25dcd6e4707bcb","internal_anchors":9},"formal_canon":{"evidence_count":2,"snapshot_sha256":"9fb0d4a413c47666296c1d5c3b50d42c6c7f8c57ba9ea500973266d9df332940"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}