{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:CVJTD6ZFOGT7DLP2O4ALRQJHSZ","short_pith_number":"pith:CVJTD6ZF","canonical_record":{"source":{"id":"2503.20020","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-03-25T19:02:56Z","cross_cats_sorted":[],"title_canon_sha256":"f7ede4ef033f12c3240aee316383c9e2014556de81935c004e69fb139527efcb","abstract_canon_sha256":"1e4a8f78a6c16ec3726ffc4ab90564a204264d864d71685c96bf27c7f81b7a55"},"schema_version":"1.0"},"canonical_sha256":"155331fb2571a7f1adfa7700b8c1279649c8d8364c9c8ce77c7ba5401a968eb2","source":{"kind":"arxiv","id":"2503.20020","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.20020","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"arxiv_version","alias_value":"2503.20020v1","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.20020","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"pith_short_12","alias_value":"CVJTD6ZFOGT7","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"pith_short_16","alias_value":"CVJTD6ZFOGT7DLP2","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"pith_short_8","alias_value":"CVJTD6ZF","created_at":"2026-07-05T10:39:30Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:CVJTD6ZFOGT7DLP2O4ALRQJHSZ","target":"record","payload":{"canonical_record":{"source":{"id":"2503.20020","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-03-25T19:02:56Z","cross_cats_sorted":[],"title_canon_sha256":"f7ede4ef033f12c3240aee316383c9e2014556de81935c004e69fb139527efcb","abstract_canon_sha256":"1e4a8f78a6c16ec3726ffc4ab90564a204264d864d71685c96bf27c7f81b7a55"},"schema_version":"1.0"},"canonical_sha256":"155331fb2571a7f1adfa7700b8c1279649c8d8364c9c8ce77c7ba5401a968eb2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:39:30.673550Z","signature_b64":"V8TVYAbFp9l8zgz3lPQA9m/NzHkmldwdNhasyN257ugXtcdhtrAkZlKYj+sq+MvlXtl4A64jS4oMWDgr27bTAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"155331fb2571a7f1adfa7700b8c1279649c8d8364c9c8ce77c7ba5401a968eb2","last_reissued_at":"2026-07-05T10:39:30.673043Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:39:30.673043Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.20020","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:39:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"O/NIr3M6+UoNiq38akW7DMxsx9pkTJcZo/4cQCUEIDYeW3zLzrJwHuBVnae63BeDz1X0nwRt3ABth1ZH/TciAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-25T09:03:47.869522Z"},"content_sha256":"7f1b8cb1c7977874e3418676a60e94c05ab7de2ec9843c7b3a9c78eb4bbbca1e","schema_version":"1.0","event_id":"sha256:7f1b8cb1c7977874e3418676a60e94c05ab7de2ec9843c7b3a9c78eb4bbbca1e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:CVJTD6ZFOGT7DLP2O4ALRQJHSZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Gemini Robotics: Bringing AI into the Physical World","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Gemini Robotics is a Vision-Language-Action model that directly controls robots to perform complex manipulation tasks in varied and unseen environments.","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Acorn Pooley, Adil Dostmohamed, Alexander Herzog, Alex X. Lee, Allan Zhou, Anirudha Majumdar, Annie Xie, Anthony Brohan, Antoine Laurens, Arunkumar Byravan, Ashwin Balakrishna, Assaf Hurwitz Michaely, Atil Iscen, Ayzaan Wahid, Brandon Hernaez, Carolina Parada, Charles Shu, Chase Kew, Chuyuan Fu, Claudio Fantacci, Cody Fong, Coline Devin, Danny Driess, David D'Ambrosio, Debidatta Dwibedi, Deepali Jain, Dhruv Shah, Dmitry Kalashnikov, Dorsa Sadigh, Emilio Parisotto, Erik Frey, Federico Casarini, Fei Xia, Francesco Nori, Gemini Robotics Team, Giulia Vezzani, Grace Vesom, Hao-Tien Lewis Chiang, Isabel Leal, Jacky Liang, Jake Varley, Jan Humplik, Jean-Baptiste Alayrac, Jerad Kirkland, Jie Tan, Jingwei Zhang, Jinyu Xie, Jonathan Tompson, Jose Enrique Chen, Joshua Ainslie, Jost Tobias Springenberg, Kanishka Rao, Kathryn Shea, Keerthana Gopalakrishnan, Ken Caluwaerts, Konstantinos Bousmalis, Krista Reymann, Krzysztof Choromanski, Laura Graesser, Leonard Hasenclever, Maria Bauza, Marissa Giustina, M. Emre Karagozler, Michael Elabd, Michael Neunert, Michiel Blokzijl, Mithun George Jacob, Mohit Sharma, Montserrat Gonzalez Arenas, Nicolas Heess, Norman Di Palo, Oriol Vinyals, Oscar Chang, Pannag Sanketi, Paul Wohlhart, Peng Xu, Peter Pastor, Pierre Sermanet, Rachel Sterneck, Radu Soricut, R. Alex Hofer, Razvan Surdulescu, Robert Baruch, Robert Moreno, Rui Yao, Ryan Julian, Saminda Abeyruwan, Sean Kirmani, Sergey Yaroshenko, Serkan Cabi, Sharath Maddineni, Sichun Xu, Stefani Karp, Stefano Saliceti, Stefan Welker, Steven Bohez, Sudeep Dasari, Sumeet Singh, Ted Xiao, Thomas Buschmann, Thomas Lampe, Tianli Ding, Tingnan Zhang, Todor Davchev, Travis Armstrong, Tsang-Wei Edward Lee, Vikas Sindhwani, Vincent Vanhoucke, Wenhao Yu, Wentao Yuan, Xi Chen, Yilun Du, Ying Xu, Yixin Lin, Yuheng Kuang, Yuxiang Yang, Yuxiang Zhou, Zhuo Xu","submitted_at":"2025-03-25T19:02:56Z","abstract_excerpt":"Recent advancements in large multimodal models have led to the emergence of remarkable generalist capabilities in digital domains, yet their translation to physical agents such as robots remains a significant challenge. This report introduces a new family of AI models purposefully designed for robotics and built upon the foundation of Gemini 2.0. We present Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots. Gemini Robotics executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots, executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being robust to variations in object types and positions, handling unseen environments as well as following diverse, open vocabulary instructions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the model's claimed robustness, generalization to unseen environments, and ability to learn from as few as 100 demonstrations will hold when deployed on actual physical robots in real-world conditions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Gemini Robotics is a Vision-Language-Action model for robot control that handles complex tasks robustly and adapts with minimal data, supported by an embodied reasoning extension.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Gemini Robotics is a Vision-Language-Action model that directly controls robots to perform complex manipulation tasks in varied and unseen environments.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1ee2083d8b802a908cfa64def8a7956b6893e529c8ff72d9941c50285f63e17f"},"source":{"id":"2503.20020","kind":"arxiv","version":1},"verdict":{"id":"f0a66a28-1448-4c99-8567-d9abe94b8f9b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-11T15:24:29.907364Z","strongest_claim":"Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots, executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being robust to variations in object types and positions, handling unseen environments as well as following diverse, open vocabulary instructions.","one_line_summary":"Gemini Robotics is a Vision-Language-Action model for robot control that handles complex tasks robustly and adapts with minimal data, supported by an embodied reasoning extension.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the model's claimed robustness, generalization to unseen environments, and ability to learn from as few as 100 demonstrations will hold when deployed on actual physical robots in real-world conditions.","pith_extraction_headline":"Gemini Robotics is a Vision-Language-Action model that directly controls robots to perform complex manipulation tasks in varied and unseen environments."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2503.20020/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":34,"sample":[{"doi":"","year":2022,"title":"Constitutional AI: Harmlessness from AI Feedback","work_id":"faaaa4e0-2676-4fac-a0b4-99aef10d2095","ref_index":1,"cited_arxiv_id":"2212.08073","is_internal_anchor":true},{"doi":"10.1109/lra.2024.3410155","year":2021,"title":"Language models as zero-shot trajectory generators","work_id":"b5dd20ba-8222-41d6-8fea-612d2427a5a7","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1109/cvpr46437.2021","year":2024,"title":"Derf: Decomposed radiance fields","work_id":"7083a41e-5666-435b-ab26-c753f6490b9a","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"Contributions and Acknowledgments Authors Saminda Abeyruwan Joshua Ainslie Jean-Baptiste Alayrac Montserrat Gonzalez Arenas Travis Armstrong Ashwin Balakrishna Robert Baruch Maria Bauza Michiel Blokzi","work_id":"8aefe41b-b604-4de0-8b4f-51dba446a078","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"The user will provide a task instruction along with an initial image of the workspace area from the overhead camera, initial robot state and initial scene objects","work_id":"fe8b4bf2-fb1e-4eeb-b15a-20c5fdb8e8da","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":34,"snapshot_sha256":"873b51b5dbd76b9fbaff3ce72bb0aa4c132e0d1868e6d3e767204459053fa22b","internal_anchors":1},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"f0a66a28-1448-4c99-8567-d9abe94b8f9b"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:39:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nEDJI3mcQszp83jsOz6hPuuUJJvrtHke1CMdayCqpM3UKrj6QnaAwAF1iPIX+tvagD5bYoqcMoFSILJEn1x3DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-25T09:03:47.870102Z"},"content_sha256":"0285247c29d9c00c14f5ba9b013a5a357df8287ff2722b50a22fb8810d744661","schema_version":"1.0","event_id":"sha256:0285247c29d9c00c14f5ba9b013a5a357df8287ff2722b50a22fb8810d744661"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ/bundle.json","state_url":"https://pith.science/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-25T09:03:47Z","links":{"resolver":"https://pith.science/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ","bundle":"https://pith.science/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ/bundle.json","state":"https://pith.science/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CVJTD6ZFOGT7DLP2O4ALRQJHSZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:CVJTD6ZFOGT7DLP2O4ALRQJHSZ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"1e4a8f78a6c16ec3726ffc4ab90564a204264d864d71685c96bf27c7f81b7a55","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-03-25T19:02:56Z","title_canon_sha256":"f7ede4ef033f12c3240aee316383c9e2014556de81935c004e69fb139527efcb"},"schema_version":"1.0","source":{"id":"2503.20020","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.20020","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"arxiv_version","alias_value":"2503.20020v1","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.20020","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"pith_short_12","alias_value":"CVJTD6ZFOGT7","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"pith_short_16","alias_value":"CVJTD6ZFOGT7DLP2","created_at":"2026-07-05T10:39:30Z"},{"alias_kind":"pith_short_8","alias_value":"CVJTD6ZF","created_at":"2026-07-05T10:39:30Z"}],"graph_snapshots":[{"event_id":"sha256:0285247c29d9c00c14f5ba9b013a5a357df8287ff2722b50a22fb8810d744661","target":"graph","created_at":"2026-07-05T10:39:30Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots, executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being robust to variations in object types and positions, handling unseen environments as well as following diverse, open vocabulary instructions."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the model's claimed robustness, generalization to unseen environments, and ability to learn from as few as 100 demonstrations will hold when deployed on actual physical robots in real-world conditions."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Gemini Robotics is a Vision-Language-Action model for robot control that handles complex tasks robustly and adapts with minimal data, supported by an embodied reasoning extension."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Gemini Robotics is a Vision-Language-Action model that directly controls robots to perform complex manipulation tasks in varied and unseen environments."}],"snapshot_sha256":"1ee2083d8b802a908cfa64def8a7956b6893e529c8ff72d9941c50285f63e17f"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2503.20020/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advancements in large multimodal models have led to the emergence of remarkable generalist capabilities in digital domains, yet their translation to physical agents such as robots remains a significant challenge. This report introduces a new family of AI models purposefully designed for robotics and built upon the foundation of Gemini 2.0. We present Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots. Gemini Robotics executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being ","authors_text":"Acorn Pooley, Adil Dostmohamed, Alexander Herzog, Alex X. Lee, Allan Zhou, Anirudha Majumdar, Annie Xie, Anthony Brohan, Antoine Laurens, Arunkumar Byravan, Ashwin Balakrishna, Assaf Hurwitz Michaely, Atil Iscen, Ayzaan Wahid, Brandon Hernaez, Carolina Parada, Charles Shu, Chase Kew, Chuyuan Fu, Claudio Fantacci, Cody Fong, Coline Devin, Danny Driess, David D'Ambrosio, Debidatta Dwibedi, Deepali Jain, Dhruv Shah, Dmitry Kalashnikov, Dorsa Sadigh, Emilio Parisotto, Erik Frey, Federico Casarini, Fei Xia, Francesco Nori, Gemini Robotics Team, Giulia Vezzani, Grace Vesom, Hao-Tien Lewis Chiang, Isabel Leal, Jacky Liang, Jake Varley, Jan Humplik, Jean-Baptiste Alayrac, Jerad Kirkland, Jie Tan, Jingwei Zhang, Jinyu Xie, Jonathan Tompson, Jose Enrique Chen, Joshua Ainslie, Jost Tobias Springenberg, Kanishka Rao, Kathryn Shea, Keerthana Gopalakrishnan, Ken Caluwaerts, Konstantinos Bousmalis, Krista Reymann, Krzysztof Choromanski, Laura Graesser, Leonard Hasenclever, Maria Bauza, Marissa Giustina, M. Emre Karagozler, Michael Elabd, Michael Neunert, Michiel Blokzijl, Mithun George Jacob, Mohit Sharma, Montserrat Gonzalez Arenas, Nicolas Heess, Norman Di Palo, Oriol Vinyals, Oscar Chang, Pannag Sanketi, Paul Wohlhart, Peng Xu, Peter Pastor, Pierre Sermanet, Rachel Sterneck, Radu Soricut, R. Alex Hofer, Razvan Surdulescu, Robert Baruch, Robert Moreno, Rui Yao, Ryan Julian, Saminda Abeyruwan, Sean Kirmani, Sergey Yaroshenko, Serkan Cabi, Sharath Maddineni, Sichun Xu, Stefani Karp, Stefano Saliceti, Stefan Welker, Steven Bohez, Sudeep Dasari, Sumeet Singh, Ted Xiao, Thomas Buschmann, Thomas Lampe, Tianli Ding, Tingnan Zhang, Todor Davchev, Travis Armstrong, Tsang-Wei Edward Lee, Vikas Sindhwani, Vincent Vanhoucke, Wenhao Yu, Wentao Yuan, Xi Chen, Yilun Du, Ying Xu, Yixin Lin, Yuheng Kuang, Yuxiang Yang, Yuxiang Zhou, Zhuo Xu","cross_cats":[],"headline":"Gemini Robotics is a Vision-Language-Action model that directly controls robots to perform complex manipulation tasks in varied and unseen environments.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-03-25T19:02:56Z","title":"Gemini Robotics: Bringing AI into the Physical World"},"references":{"count":34,"internal_anchors":1,"resolved_work":34,"sample":[{"cited_arxiv_id":"2212.08073","doi":"","is_internal_anchor":true,"ref_index":1,"title":"Constitutional AI: Harmlessness from AI Feedback","work_id":"faaaa4e0-2676-4fac-a0b4-99aef10d2095","year":2022},{"cited_arxiv_id":"","doi":"10.1109/lra.2024.3410155","is_internal_anchor":false,"ref_index":2,"title":"Language models as zero-shot trajectory generators","work_id":"b5dd20ba-8222-41d6-8fea-612d2427a5a7","year":2021},{"cited_arxiv_id":"","doi":"10.1109/cvpr46437.2021","is_internal_anchor":false,"ref_index":3,"title":"Derf: Decomposed radiance fields","work_id":"7083a41e-5666-435b-ab26-c753f6490b9a","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Contributions and Acknowledgments Authors Saminda Abeyruwan Joshua Ainslie Jean-Baptiste Alayrac Montserrat Gonzalez Arenas Travis Armstrong Ashwin Balakrishna Robert Baruch Maria Bauza Michiel Blokzi","work_id":"8aefe41b-b604-4de0-8b4f-51dba446a078","year":2019},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"The user will provide a task instruction along with an initial image of the workspace area from the overhead camera, initial robot state and initial scene objects","work_id":"fe8b4bf2-fb1e-4eeb-b15a-20c5fdb8e8da","year":null}],"snapshot_sha256":"873b51b5dbd76b9fbaff3ce72bb0aa4c132e0d1868e6d3e767204459053fa22b"},"source":{"id":"2503.20020","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-11T15:24:29.907364Z","id":"f0a66a28-1448-4c99-8567-d9abe94b8f9b","model_set":{"reader":"grok-4.3"},"one_line_summary":"Gemini Robotics is a Vision-Language-Action model for robot control that handles complex tasks robustly and adapts with minimal data, supported by an embodied reasoning extension.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Gemini Robotics is a Vision-Language-Action model that directly controls robots to perform complex manipulation tasks in varied and unseen environments.","strongest_claim":"Gemini Robotics, an advanced Vision-Language-Action (VLA) generalist model capable of directly controlling robots, executes smooth and reactive movements to tackle a wide range of complex manipulation tasks while also being robust to variations in object types and positions, handling unseen environments as well as following diverse, open vocabulary instructions.","weakest_assumption":"That the model's claimed robustness, generalization to unseen environments, and ability to learn from as few as 100 demonstrations will hold when deployed on actual physical robots in real-world conditions."}},"verdict_id":"f0a66a28-1448-4c99-8567-d9abe94b8f9b"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7f1b8cb1c7977874e3418676a60e94c05ab7de2ec9843c7b3a9c78eb4bbbca1e","target":"record","created_at":"2026-07-05T10:39:30Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"1e4a8f78a6c16ec3726ffc4ab90564a204264d864d71685c96bf27c7f81b7a55","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-03-25T19:02:56Z","title_canon_sha256":"f7ede4ef033f12c3240aee316383c9e2014556de81935c004e69fb139527efcb"},"schema_version":"1.0","source":{"id":"2503.20020","kind":"arxiv","version":1}},"canonical_sha256":"155331fb2571a7f1adfa7700b8c1279649c8d8364c9c8ce77c7ba5401a968eb2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"155331fb2571a7f1adfa7700b8c1279649c8d8364c9c8ce77c7ba5401a968eb2","first_computed_at":"2026-07-05T10:39:30.673043Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:39:30.673043Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"V8TVYAbFp9l8zgz3lPQA9m/NzHkmldwdNhasyN257ugXtcdhtrAkZlKYj+sq+MvlXtl4A64jS4oMWDgr27bTAA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:39:30.673550Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.20020","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7f1b8cb1c7977874e3418676a60e94c05ab7de2ec9843c7b3a9c78eb4bbbca1e","sha256:0285247c29d9c00c14f5ba9b013a5a357df8287ff2722b50a22fb8810d744661"],"state_sha256":"4cc5306c65c7e94de84ae2b79403ebe0b850446b490866e338bf7f0dc66d18e0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+MdISf+Qywa0LdltKiXe6mZKwNJKSgb1B29iYhnsBRwWhTZbk8jGZjAtw578BfzZZH0ohZ8ShwPpGtey1LsPBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-25T09:03:47.873098Z","bundle_sha256":"d6db78e57ec9ddbfa8f934e7135c0018fc5d564811ea4dc4c5edf569aff4af1c"}}