{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/emerging-properties-in-unified-multimodal","title":"Emerging Properties in Unified Multimodal Pretraining","arxiv_id":"2505.14683","date":"2025-05-20","proceeding":null,"authors":["Chaorui Deng","Deyao Zhu","Kunchang Li","Chenhui Gou","Feng Li","Zeyu Wang","Shu Zhong","Weihao Yu","Xiaonan Nie","Ziang Song","Guang Shi","Haoqi Fan"],"abstract":"Unifying multimodal understanding and generation has shown impressive capabilities in cutting-edge proprietary systems. In this work, we introduce BAGEL, an open0source foundational model that natively supports multimodal understanding and generation. BAGEL is a unified, decoder0only model pretrained on trillions of tokens curated from large0scale interleaved text, image, video, and web data. When scaled with such diverse multimodal interleaved data, BAGEL exhibits emerging capabilities in complex multimodal reasoning. As a result, it significantly outperforms open-source unified models in both multimodal generation and understanding across standard benchmarks, while exhibiting advanced multimodal reasoning abilities such as free-form image manipulation, future frame prediction, 3D manipulation, and world navigation. In the hope of facilitating further opportunities for multimodal research, we share the key findings, pretraining details, data creation protocal, and release our code and checkpoints to the community. The project page is at https://bagel-ai.org/","url_abs":"https://arxiv.org/abs/2505.14683v1","url_pdf":"https://arxiv.org/pdf/2505.14683v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"emerging-properties-in-unified-multimodal","repo_url":"https://github.com/ByteDance-Seed/Bagel","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"emerging-properties-in-unified-multimodal","repo_url":"https://github.com/neverbiasu/ComfyUI-BAGEL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"image-editing","task_name":"Image Editing"},{"task_slug":"image-generation","task_name":"Image Generation"},{"task_slug":"image-manipulation","task_name":"Image Manipulation"},{"task_slug":"multimodal-reasoning","task_name":"Multimodal Reasoning"},{"task_slug":"multimodal-generation","task_name":"multimodal generation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/on-wise","task":"","dataset":"WISE","model":"BAGEL","rank_in_archive_order":3,"of":11,"metrics":{"Biology":"0.44","Chemistry":"0.39","Cultural":"0.44","Overall":"0.52","Physics":"0.60","Space":"0.68","Time":"0.55"},"uses_additional_data":false},{"leaderboard":"/sota/image-editing-on-gedit-bench-en","task":"Image Editing","dataset":"GEdit-Bench-EN","model":"BAGEL","rank_in_archive_order":3,"of":3,"metrics":{"Overall":"6.52","Perceptual Quality":"6.83","Semantic Consistency":"7.36"},"uses_additional_data":true},{"leaderboard":"/sota/image-editing-on-imgedit-data","task":"Image Editing","dataset":"ImgEdit-Data","model":"BAGEL","rank_in_archive_order":3,"of":9,"metrics":{"Action":"4.17","Add":"3.56","Adjust":"3.31","Background":"3.24","Extract":"1.70","Hybrid":"2.38","Overall":"3.20","Remove":"2.62","Replace":"3.30","Style":"4.49"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-wise","task":"Image Generation","dataset":"WISE","model":"Bagel (w/ cot)","rank_in_archive_order":2,"of":14,"metrics":{"Biology":"0.65","Chemistry":"0.58","Cultural":"0.76","Overall":"0.70","Physics":"0.75","Space":"0.75","Time":"0.69"},"uses_additional_data":false},{"leaderboard":"/sota/image-generation-on-wise","task":"Image Generation","dataset":"WISE","model":"Bagel","rank_in_archive_order":5,"of":14,"metrics":{"Biology":"0.44","Chemistry":"0.39","Cultural":"0.44","Overall":"0.52","Physics":"0.60","Space":"0.68","Time":"0.55"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2505.14683","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2505.14683"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/neverbiasu/ComfyUI-BAGEL","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ByteDance-Seed/Bagel","reach":null}],"summary":{"ran":3,"ran_honours":3,"ran_fixture":2,"ran_draft_wrong":1,"unverified":12},"by_repo_kind":{"listed":{"samples":21,"ran":9,"repositories":2}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"4102c144078044c7","entry":"apply_rotary_pos_emb","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/siglip_navit.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/siglip_navit.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"4102c144078044c7"}},{"code_sha256_prefix":"fae038c6ae265486","entry":"count_parameters","repo":"ByteDance-Seed/Bagel","repo_kind":"listed","path":"train/pretrain_unified_navit.py","file_url":"https://github.com/ByteDance-Seed/Bagel/blob/HEAD/train/pretrain_unified_navit.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"fae038c6ae265486"}},{"code_sha256_prefix":"5bd79f64dffa39d7","entry":"detect_peak_tflops","repo":"ByteDance-Seed/Bagel","repo_kind":"listed","path":"train/pretrain_unified_navit.py","file_url":"https://github.com/ByteDance-Seed/Bagel/blob/HEAD/train/pretrain_unified_navit.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"5bd79f64dffa39d7"}},{"code_sha256_prefix":"e5947aba1d10885f","entry":"get_1d_sincos_pos_embed_from_grid","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/modeling_utils.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/modeling_utils.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"e5947aba1d10885f"}},{"code_sha256_prefix":"58f584dfd7b8fc3f","entry":"get_2d_sincos_pos_embed","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/modeling_utils.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/modeling_utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"58f584dfd7b8fc3f"}},{"code_sha256_prefix":"26da3ed73c66b411","entry":"pad_sequence","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/qwen2_navit.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/qwen2_navit.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"26da3ed73c66b411"}},{"code_sha256_prefix":"eeb5aa9c76e8e1ae","entry":"qwen2_flop_coefficients","repo":"ByteDance-Seed/Bagel","repo_kind":"listed","path":"train/pretrain_unified_navit.py","file_url":"https://github.com/ByteDance-Seed/Bagel/blob/HEAD/train/pretrain_unified_navit.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"eeb5aa9c76e8e1ae"}},{"code_sha256_prefix":"cec833a99feb528c","entry":"rotate_half","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/siglip_navit.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/siglip_navit.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"cec833a99feb528c"}},{"code_sha256_prefix":"844a4aecea80d106","entry":"timed","repo":"ByteDance-Seed/Bagel","repo_kind":"listed","path":"eval/gen/geneval/evaluation/evaluate_images.py","file_url":"https://github.com/ByteDance-Seed/Bagel/blob/HEAD/eval/gen/geneval/evaluation/evaluate_images.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"844a4aecea80d106"}},{"code_sha256_prefix":"94a6a40b906a27a2","entry":"build_new_token_ids","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/runtime.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/runtime.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"94a6a40b906a27a2"}},{"code_sha256_prefix":"d37d55803e67b95d","entry":"comfy_raw_to_bagel_scaled","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/latent.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/latent.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"d37d55803e67b95d"}},{"code_sha256_prefix":"446c7a69810151e0","entry":"discover_converted_bagel","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/discovery.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/discovery.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"446c7a69810151e0"}},{"code_sha256_prefix":"908a8209ec674ee6","entry":"flash_attn_varlen_func","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/flash_attention.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/flash_attention.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"908a8209ec674ee6"}},{"code_sha256_prefix":"f9f0a4254ea96f69","entry":"from_comfy_latent","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/latent.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/latent.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f9f0a4254ea96f69"}},{"code_sha256_prefix":"665d8a4e8f673a4c","entry":"get_2d_sincos_pos_embed_from_grid","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/modeling_utils.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/modeling_utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"665d8a4e8f673a4c"}},{"code_sha256_prefix":"1db1e937be531f0d","entry":"normalize_key","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/converted_format.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/converted_format.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"1db1e937be531f0d"}},{"code_sha256_prefix":"26c4c48bbb6ab404","entry":"normalize_state_dict","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/converted_format.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/converted_format.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"26c4c48bbb6ab404"}},{"code_sha256_prefix":"1a10ae6dd1f23878","entry":"reload_bagel_patcher","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/model_patcher.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/model_patcher.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"1a10ae6dd1f23878"}},{"code_sha256_prefix":"7f5ffed1f98ca2b1","entry":"sha256_bytes","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/converted_format.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/converted_format.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"7f5ffed1f98ca2b1"}},{"code_sha256_prefix":"312ebc115176cd64","entry":"to_comfy_latent","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/latent.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/latent.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"312ebc115176cd64"}},{"code_sha256_prefix":"0b23c6505f5db1d0","entry":"update_context_text","repo":"neverbiasu/ComfyUI-BAGEL","repo_kind":"listed","path":"modeling/bagel/runtime.py","file_url":"https://github.com/neverbiasu/ComfyUI-BAGEL/blob/HEAD/modeling/bagel/runtime.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0b23c6505f5db1d0"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}