{"work":{"id":"c8b07deb-8fe7-4e18-9620-f3569d3529ce","openalex_id":"https://openalex.org/W4405356851","doi":"10.1055/a-2487-1252","arxiv_id":"2508.10104","raw_key":null,"title":"DINOv3","authors":null,"authors_text":"Oriane Siméoni, Huy V","year":2025,"venue":"cs.CV","abstract":"Self-supervised learning holds the promise of eliminating the need for manual data annotation, enabling models to scale effortlessly to massive datasets and larger architectures. By not being tailored to specific tasks or domains, this training paradigm has the potential to learn visual representations from diverse sources, ranging from natural to aerial images -- using a single algorithm. This technical report introduces DINOv3, a major milestone toward realizing this vision by leveraging simple yet effective strategies. First, we leverage the benefit of scaling both dataset and model size by careful data preparation, design, and optimization. Second, we introduce a new method called Gram anchoring, which effectively addresses the known yet unsolved issue of dense feature maps degrading during long training schedules. Finally, we apply post-hoc strategies that further enhance our models' flexibility with respect to resolution, model size, and alignment with text. As a result, we present a versatile vision foundation model that outperforms the specialized state of the art across a broad range of settings, without fine-tuning. DINOv3 produces high-quality dense features that achieve outstanding performance on various vision tasks, significantly surpassing previous self- and weakly-supervised foundation models. We also share the DINOv3 suite of vision models, designed to advance the state of the art on a wide spectrum of tasks and data by providing scalable solutions for diverse resource constraints and deployment scenarios.","external_url":"https://arxiv.org/abs/2508.10104","cited_by_count":4,"metadata_source":"pith","metadata_fetched_at":"2026-08-05T02:28:24.338817+00:00","pith_arxiv_id":"2508.10104","created_at":"2026-05-08T21:09:12.812513+00:00","updated_at":"2026-08-05T02:28:24.338817+00:00","title_quality_ok":false,"display_title":"DINOv3","render_title":"DINOv3"},"hub":{"state":{"work_id":"c8b07deb-8fe7-4e18-9620-f3569d3529ce","tier":"super_hub","tier_reason":"100+ Pith inbound or 10,000+ external citations","pith_inbound_count":597,"external_cited_by_count":4,"distinct_field_count":18,"first_pith_cited_at":"2024-11-28T12:13:46+00:00","last_pith_cited_at":"2026-07-09T14:35:37+00:00","author_build_status":"needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"needed","reader_status":"not_needed","recognition_status":"not_needed","updated_at":"2026-08-21T21:19:17.011088+00:00","tier_text":"super_hub"},"tier":"super_hub","role_counts":[{"context_role":"method","n":38},{"context_role":"background","n":37},{"context_role":"baseline","n":9},{"context_role":"dataset","n":4},{"context_role":"other","n":1}],"polarity_counts":[{"context_polarity":"background","n":38},{"context_polarity":"use_method","n":37},{"context_polarity":"baseline","n":9},{"context_polarity":"use_dataset","n":3},{"context_polarity":"extend","n":1},{"context_polarity":"unclear","n":1}],"runs":{"ask_index":{"job_type":"ask_index","status":"succeeded","result":{"title":"DINOv3","claims":[{"claim_text":"Self-supervised learning holds the promise of eliminating the need for manual data annotation, enabling models to scale effortlessly to massive datasets and larger architectures. By not being tailored to specific tasks or domains, this training paradigm has the potential to learn visual representations from diverse sources, ranging from natural to aerial images -- using a single algorithm. This technical report introduces DINOv3, a major milestone toward realizing this vision by leveraging simple yet effective strategies. First, we leverage the benefit of scaling both dataset and model size by","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DINOv3 because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T18:39:21.316713+00:00"},"author_expand":{"job_type":"author_expand","status":"succeeded","result":{"authors_linked":[{"id":"c2a08797-7d6a-4291-9163-79c500867f06","orcid":null,"display_name":"Oriane Siméoni"},{"id":"55300c1b-27d7-46cb-ad86-cc80cc68f865","orcid":null,"display_name":"Huy V"}]},"error":null,"updated_at":"2026-05-14T18:39:39.430643+00:00"},"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-14T18:36:31.420749+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"DINOv2: Learning Robust Visual Features without Supervision","work_id":"26b304e5-b54a-4f26-be7e-83299eca52e4","shared_citers":63},{"title":"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale","work_id":"e96730e3-129b-4db6-b981-15ab7932e297","shared_citers":24},{"title":"SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features","work_id":"50eec732-2d41-432f-9dcf-ac7fff235ea5","shared_citers":23},{"title":"Decoupled Weight Decay Regularization","work_id":"07ef7360-d385-4033-83f7-8384a6325204","shared_citers":21},{"title":"SAM 2: Segment Anything in Images and Videos","work_id":"acc13f66-d814-44f9-9688-375688bf2d4a","shared_citers":20},{"title":"SAM 3: Segment Anything with Concepts","work_id":"4a72a006-2592-4554-aad0-a9c41a9f952d","shared_citers":19},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":13},{"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","shared_citers":13},{"title":"Wan: Open and Advanced Large-Scale Video Generative Models","work_id":"ad3ebc3b-4224-46c9-b61d-bcf135da0a7c","shared_citers":13},{"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","shared_citers":12},{"title":"Representation Learning with Contrastive Predictive Coding","work_id":"7b08a1d4-d565-424e-9c86-6ef244b7b90a","shared_citers":12},{"title":"Adam: A Method for Stochastic Optimization","work_id":"1910796d-9b52-4683-bf5c-de9632c1028b","shared_citers":11},{"title":"Auto-Encoding Variational Bayes","work_id":"97d95295-30e1-42b4-bbf6-85f0fa4edb44","shared_citers":10},{"title":"iBOT: Image BERT Pre-Training with Online Tokenizer","work_id":"ddf5ecec-b36e-4d27-96b7-5aefc236d17c","shared_citers":10},{"title":"V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning","work_id":"a9c28401-f16a-4933-89f0-788e2f94e52b","shared_citers":10},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":9},{"title":"$\\pi_0$: A Vision-Language-Action Flow Model for General Robot Control","work_id":"f790abdc-a796-482f-a40d-f8ee035ecfc2","shared_citers":8},{"title":"BEiT: BERT Pre-Training of Image Transformers","work_id":"d74eda3c-bf7e-45f1-a8f1-a0137ecca3f4","shared_citers":8},{"title":"Flow Matching for Generative Modeling","work_id":"6edb71c4-5d64-40af-a394-9757ea051a36","shared_citers":8},{"title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","work_id":"008df105-2fdd-45d8-857a-8e35868aecb6","shared_citers":8},{"title":"Vision Transformers Need Registers","work_id":"57106da4-5420-4778-94eb-e821589aa7a0","shared_citers":8},{"title":"$\\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization","work_id":"d1ad7304-d09a-49bc-809e-846439f6aff9","shared_citers":7},{"title":"Depth Anything 3: Recovering the Visual Space from Any Views","work_id":"0a54b500-1e9d-46c2-85eb-8e16cbac8461","shared_citers":7},{"title":"Distilling the Knowledge in a Neural Network","work_id":"d927ab1f-17b8-4002-9d09-c3d55764fbad","shared_citers":7}],"time_series":[{"n":1,"year":2025},{"n":207,"year":2026}],"dependency_candidates":[]},"error":null,"updated_at":"2026-05-14T18:39:21.310487+00:00"},"identity_refresh":{"job_type":"identity_refresh","status":"succeeded","result":{"items":[{"title":"Qwen3 Technical Report","outcome":"unchanged","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","resolver":"local_arxiv","confidence":0.98,"old_work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e"}],"counts":{"fixed":0,"merged":0,"unchanged":1,"quarantined":0,"needs_external_resolution":0},"errors":[],"attempted":1},"error":null,"updated_at":"2026-05-14T18:36:37.641811+00:00"},"role_polarity":{"job_type":"role_polarity","status":"succeeded","result":{"title":"DINOv3","claims":[{"claim_text":"Self-supervised learning holds the promise of eliminating the need for manual data annotation, enabling models to scale effortlessly to massive datasets and larger architectures. By not being tailored to specific tasks or domains, this training paradigm has the potential to learn visual representations from diverse sources, ranging from natural to aerial images -- using a single algorithm. This technical report introduces DINOv3, a major milestone toward realizing this vision by leveraging simple yet effective strategies. First, we leverage the benefit of scaling both dataset and model size by","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DINOv3 because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T18:39:27.794726+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"DINOv3","claims":[{"claim_text":"Self-supervised learning holds the promise of eliminating the need for manual data annotation, enabling models to scale effortlessly to massive datasets and larger architectures. By not being tailored to specific tasks or domains, this training paradigm has the potential to learn visual representations from diverse sources, ranging from natural to aerial images -- using a single algorithm. This technical report introduces DINOv3, a major milestone toward realizing this vision by leveraging simple yet effective strategies. First, we leverage the benefit of scaling both dataset and model size by","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DINOv3 because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T18:39:32.452967+00:00"}},"summary":{"title":"DINOv3","claims":[{"claim_text":"Self-supervised learning holds the promise of eliminating the need for manual data annotation, enabling models to scale effortlessly to massive datasets and larger architectures. By not being tailored to specific tasks or domains, this training paradigm has the potential to learn visual representations from diverse sources, ranging from natural to aerial images -- using a single algorithm. This technical report introduces DINOv3, a major milestone toward realizing this vision by leveraging simple yet effective strategies. First, we leverage the benefit of scaling both dataset and model size by","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks DINOv3 because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"DINOv2: Learning Robust Visual Features without Supervision","work_id":"26b304e5-b54a-4f26-be7e-83299eca52e4","shared_citers":63},{"title":"An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale","work_id":"e96730e3-129b-4db6-b981-15ab7932e297","shared_citers":24},{"title":"SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features","work_id":"50eec732-2d41-432f-9dcf-ac7fff235ea5","shared_citers":23},{"title":"Decoupled Weight Decay Regularization","work_id":"07ef7360-d385-4033-83f7-8384a6325204","shared_citers":21},{"title":"SAM 2: Segment Anything in Images and Videos","work_id":"acc13f66-d814-44f9-9688-375688bf2d4a","shared_citers":20},{"title":"SAM 3: Segment Anything with Concepts","work_id":"4a72a006-2592-4554-aad0-a9c41a9f952d","shared_citers":19},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":13},{"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","shared_citers":13},{"title":"Wan: Open and Advanced Large-Scale Video Generative Models","work_id":"ad3ebc3b-4224-46c9-b61d-bcf135da0a7c","shared_citers":13},{"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","shared_citers":12},{"title":"Representation Learning with Contrastive Predictive Coding","work_id":"7b08a1d4-d565-424e-9c86-6ef244b7b90a","shared_citers":12},{"title":"Adam: A Method for Stochastic Optimization","work_id":"1910796d-9b52-4683-bf5c-de9632c1028b","shared_citers":11},{"title":"Auto-Encoding Variational Bayes","work_id":"97d95295-30e1-42b4-bbf6-85f0fa4edb44","shared_citers":10},{"title":"iBOT: Image BERT Pre-Training with Online Tokenizer","work_id":"ddf5ecec-b36e-4d27-96b7-5aefc236d17c","shared_citers":10},{"title":"V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning","work_id":"a9c28401-f16a-4933-89f0-788e2f94e52b","shared_citers":10},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":9},{"title":"$\\pi_0$: A Vision-Language-Action Flow Model for General Robot Control","work_id":"f790abdc-a796-482f-a40d-f8ee035ecfc2","shared_citers":8},{"title":"BEiT: BERT Pre-Training of Image Transformers","work_id":"d74eda3c-bf7e-45f1-a8f1-a0137ecca3f4","shared_citers":8},{"title":"Flow Matching for Generative Modeling","work_id":"6edb71c4-5d64-40af-a394-9757ea051a36","shared_citers":8},{"title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","work_id":"008df105-2fdd-45d8-857a-8e35868aecb6","shared_citers":8},{"title":"Vision Transformers Need Registers","work_id":"57106da4-5420-4778-94eb-e821589aa7a0","shared_citers":8},{"title":"$\\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization","work_id":"d1ad7304-d09a-49bc-809e-846439f6aff9","shared_citers":7},{"title":"Depth Anything 3: Recovering the Visual Space from Any Views","work_id":"0a54b500-1e9d-46c2-85eb-8e16cbac8461","shared_citers":7},{"title":"Distilling the Knowledge in a Neural Network","work_id":"d927ab1f-17b8-4002-9d09-c3d55764fbad","shared_citers":7}],"time_series":[{"n":1,"year":2025},{"n":207,"year":2026}],"dependency_candidates":[]},"authors":[{"id":"55300c1b-27d7-46cb-ad86-cc80cc68f865","orcid":null,"display_name":"Huy V","source":"manual","import_confidence":0.72},{"id":"c2a08797-7d6a-4291-9163-79c500867f06","orcid":null,"display_name":"Oriane Siméoni","source":"manual","import_confidence":0.72}]}}