{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GWIW4SQLWFEYCXYPQ6KLBDKVCO","short_pith_number":"pith:GWIW4SQL","schema_version":"1.0","canonical_sha256":"35916e4a0bb149815f0f8794b08d551396d8b0d459b3eb5229527201c41b28da","source":{"kind":"arxiv","id":"2506.03569","version":1},"attestation_state":"computed","paper":{"title":"MiMo-VL Technical Report","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bingquan Xia, Bowen Shen, Bowen Ye, Can Cai, Chang Liu, Chenhong He, Chong Ma, Cici, Dawei Zhu, Di Zhang, Dong Zhang, Duo Zhang, Gang Wang, Guoan Wang, Hailin Zhang, Hanglong Lv, Haochen Zhao, Hao Tian, Heng Qu, Hongshen Xu, Huaqiu Liu, Jiaming Xu, Jiebao Xiao, Jinhao Dong, Jun Shi, Kai Fang, Kainan Bao, Kangyang Zhou, Kang Zhou, Lei Li, Liang Zhao, Menghang Zhu, Nuo Chen, Peidian Li, Peng Wang, Qiantong Wang, Rui Ma, Shande Wang, Shaohui Liu, Shicheng Li, Shihua Yu, Shimao Chen, Shuhao Gu, Shuhuai Ren, Shuo Liu, Sirui Deng, Weiji Zhuang, Weikun Wang, Wei Liu, Weiwei Lv, Wenhan Ma, Wenyu Yang, Xiangwei Deng, Xiaomi LLM-Core Team: Zihao Yue, Xingchen Song, Xing Yong, Xing Zhang, Xin Zhang, Xinzhe Xu, Xu Wang, Yifan Song, Yihan Yan, Yi Huang, Yuanyuan Tian, Yudong Wang, Yue Yu, Yu Tu, Zhenbo Luo, Zhenru Lin, Zhichao Song, Zhixian Zheng, Zihan Jiang, Zihan Zhang","submitted_at":"2025-06-04T04:32:54Z","abstract_excerpt":"We open-source MiMo-VL-7B-SFT and MiMo-VL-7B-RL, two powerful vision-language models delivering state-of-the-art performance in both general visual understanding and multimodal reasoning. MiMo-VL-7B-RL outperforms Qwen2.5-VL-7B on 35 out of 40 evaluated tasks, and scores 59.4 on OlympiadBench, surpassing models with up to 78B parameters. For GUI grounding applications, it sets a new standard with 56.1 on OSWorld-G, even outperforming specialized models such as UI-TARS. Our training combines four-stage pre-training (2.4 trillion tokens) with Mixed On-policy Reinforcement Learning (MORL) integra"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.03569","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-04T04:32:54Z","cross_cats_sorted":[],"title_canon_sha256":"150696a54f0f3a061dc6f54d94f97b5425515d3b9116d6ec421aa31846d5284e","abstract_canon_sha256":"ccc66ff0127bf433fd200960d75917ac8c9fbac874f1ab81abcebcd876efc89a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:15:30.408230Z","signature_b64":"64gmn3gvN84ZLbdu1nX8Dq+URh2F1RXxL6did3M72xF+N6mgNrlq8vpQuWZ73xZo0Sl1sUKaLFVsJOIclwpkBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"35916e4a0bb149815f0f8794b08d551396d8b0d459b3eb5229527201c41b28da","last_reissued_at":"2026-07-05T11:15:30.407694Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:15:30.407694Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MiMo-VL Technical Report","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bingquan Xia, Bowen Shen, Bowen Ye, Can Cai, Chang Liu, Chenhong He, Chong Ma, Cici, Dawei Zhu, Di Zhang, Dong Zhang, Duo Zhang, Gang Wang, Guoan Wang, Hailin Zhang, Hanglong Lv, Haochen Zhao, Hao Tian, Heng Qu, Hongshen Xu, Huaqiu Liu, Jiaming Xu, Jiebao Xiao, Jinhao Dong, Jun Shi, Kai Fang, Kainan Bao, Kangyang Zhou, Kang Zhou, Lei Li, Liang Zhao, Menghang Zhu, Nuo Chen, Peidian Li, Peng Wang, Qiantong Wang, Rui Ma, Shande Wang, Shaohui Liu, Shicheng Li, Shihua Yu, Shimao Chen, Shuhao Gu, Shuhuai Ren, Shuo Liu, Sirui Deng, Weiji Zhuang, Weikun Wang, Wei Liu, Weiwei Lv, Wenhan Ma, Wenyu Yang, Xiangwei Deng, Xiaomi LLM-Core Team: Zihao Yue, Xingchen Song, Xing Yong, Xing Zhang, Xin Zhang, Xinzhe Xu, Xu Wang, Yifan Song, Yihan Yan, Yi Huang, Yuanyuan Tian, Yudong Wang, Yue Yu, Yu Tu, Zhenbo Luo, Zhenru Lin, Zhichao Song, Zhixian Zheng, Zihan Jiang, Zihan Zhang","submitted_at":"2025-06-04T04:32:54Z","abstract_excerpt":"We open-source MiMo-VL-7B-SFT and MiMo-VL-7B-RL, two powerful vision-language models delivering state-of-the-art performance in both general visual understanding and multimodal reasoning. MiMo-VL-7B-RL outperforms Qwen2.5-VL-7B on 35 out of 40 evaluated tasks, and scores 59.4 on OlympiadBench, surpassing models with up to 78B parameters. For GUI grounding applications, it sets a new standard with 56.1 on OSWorld-G, even outperforming specialized models such as UI-TARS. Our training combines four-stage pre-training (2.4 trillion tokens) with Mixed On-policy Reinforcement Learning (MORL) integra"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.03569","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2506.03569/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.03569","created_at":"2026-07-05T11:15:30.407756+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.03569v1","created_at":"2026-07-05T11:15:30.407756+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.03569","created_at":"2026-07-05T11:15:30.407756+00:00"},{"alias_kind":"pith_short_12","alias_value":"GWIW4SQLWFEY","created_at":"2026-07-05T11:15:30.407756+00:00"},{"alias_kind":"pith_short_16","alias_value":"GWIW4SQLWFEYCXYP","created_at":"2026-07-05T11:15:30.407756+00:00"},{"alias_kind":"pith_short_8","alias_value":"GWIW4SQL","created_at":"2026-07-05T11:15:30.407756+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":38,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24233","citing_title":"Latent Visual States for Efficient Multimodal Reasoning","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23678","citing_title":"AIR: Adaptive Interleaved Reasoning with Code in MLLMs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2607.02096","citing_title":"LongEgoRefer: A Benchmark for Long-Form Egocentric Video Referring Expression Comprehension","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12125","citing_title":"Q-Fold: Query-Aware Focus-Context Spatio-Temporal Folding for Long Video Understanding","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08239","citing_title":"When No Answer Is Correct: Diagnosing Absent Answer Detection for MLLMs in Video Understanding","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28551","citing_title":"DataComp-VLM: Improved Open Datasets for Vision-Language Models","ref_index":277,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04591","citing_title":"Fine-grained Fragment Retrieval in Multi-modal Long-form Dialogues","ref_index":136,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03920","citing_title":"Benchmarking Visual State Tracking in Multimodal Video Understanding","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01599","citing_title":"TRON: Targeted Rule-Verifiable Online Environments for Visual Reasoning RL","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27500","citing_title":"Aloe-Vision: Robust Vision-Language Models for Healthcare","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28551","citing_title":"DataComp-VLM: Improved Open Datasets for Vision-Language Models","ref_index":277,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18740","citing_title":"Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11651","citing_title":"Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14733","citing_title":"Video-Zero: Self-Evolution Video Understanding","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24503","citing_title":"FoodMonitor: Benchmarking MLLMs for Explainable Compliance Analysis","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25706","citing_title":"Towards Open-World Referring Expression Comprehension: A Benchmark with Training-free Multi-task Consistency Checker","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.31174","citing_title":"Detect in Any Scene: An Agentic Framework for Object Detection with Experience-Aware Reasoning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2504.12501","citing_title":"Reinforcement Learning from Human Feedback","ref_index":174,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21973","citing_title":"Foresee-to-Ground: From Predictive Temporal Perception to Evidence-Driven Reasoning for Video Temporal Grounding","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21954","citing_title":"MLLMs Know When Before Speaking: Revealing and Recovering Temporal Grounding via Attention Cues","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2602.04802","citing_title":"VISTA-Bench: Do Vision-Language Models Really Understand Visualized Text as Well as Pure Text?","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04128","citing_title":"JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation","ref_index":70,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18740","citing_title":"Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11651","citing_title":"Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2508.09521","citing_title":"PEER: Unified Process-Outcome Reinforcement Learning for Structured Empathetic Reasoning","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO","json":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO.json","graph_json":"https://pith.science/api/pith-number/GWIW4SQLWFEYCXYPQ6KLBDKVCO/graph.json","events_json":"https://pith.science/api/pith-number/GWIW4SQLWFEYCXYPQ6KLBDKVCO/events.json","paper":"https://pith.science/paper/GWIW4SQL"},"agent_actions":{"view_html":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO","download_json":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO.json","view_paper":"https://pith.science/paper/GWIW4SQL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.03569&json=true","fetch_graph":"https://pith.science/api/pith-number/GWIW4SQLWFEYCXYPQ6KLBDKVCO/graph.json","fetch_events":"https://pith.science/api/pith-number/GWIW4SQLWFEYCXYPQ6KLBDKVCO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO/action/storage_attestation","attest_author":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO/action/author_attestation","sign_citation":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO/action/citation_signature","submit_replication":"https://pith.science/pith/GWIW4SQLWFEYCXYPQ6KLBDKVCO/action/replication_record"}},"created_at":"2026-07-05T11:15:30.407756+00:00","updated_at":"2026-07-05T11:15:30.407756+00:00"}