# Thaddäus Wiedemer and coauthors argue in Google DeepMind’s paper that video models can reason [Canonical reading](https://escapemeridian.com/readings/thaddaus-wiedemer-and-coauthors-argue-in-google-deepmind-s-5ea5460e) The paper was published on September 24, 2025. Thaddäus Wiedemer and coauthors report that Veo 3 can perform tasks it was not explicitly trained for, including segmenting objects, detecting edges, editing images and understanding physical properties. They say these abilities enable early visual reasoning, such as maze and symmetry solving. The authors draw a parallel between the simple primitives behind Large Language Models’ general-purpose language abilities—large generative models trained on web-scale data—and those used by generative video models. They argue that Veo’s reported zero-shot capabilities indicate video models are on a path toward unified, generalist vision foundation models. The paper’s argument matters because it presents video models as potentially able to understand and manipulate the visual world across tasks, not only generate video. The abstract does not detail how the capabilities were evaluated or establish how well they work beyond the reported examples. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source Google DeepMind · Article Original source date: Sep 24, 2025 (2025-09-24T00:00:00+00:00). Date basis: Date stated on the publisher’s page. Added to Escape Meridian: 2026-11-12T03:43:26.106Z. First observed: 2026-10-29T01:09:48.146Z. Retrieved: 2026-10-29T01:11:40.256Z. - [Video models are zero-shot learners and reasoners](https://deepmind.google/research/publications/203190/) ## Limits The abstract does not detail how the reported capabilities were evaluated or how well they work beyond the listed tasks. Text as retrieved on October 29, 2026; originally published September 24, 2025. ## Open question What evidence would show whether the reported task abilities amount to general-purpose vision understanding? ## Source-stated Wiedemer and coauthors report that Veo 3 performs a broad range of tasks it was not explicitly trained for, including visual perception and editing tasks. > We demonstrate that Veo 3 can solve a broad variety of tasks it wasn't explicitly trained for: segmenting objects, detecting edges, editing images, understanding physical properties - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source-stated The authors say these abilities enable early forms of visual reasoning, including solving mazes and symmetries. > These abilities to perceive, model, and manipulate the visual world enable early forms of visual reasoning like maze and symmetry solving. - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source-stated The authors argue that Veo's emergent zero-shot capabilities indicate that video models are on a path toward unified, generalist vision foundation models. > Veo's emergent zero-shot capabilities indicate that video models are on a path to becoming unified, generalist vision foundation models. - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source-stated The authors draw a parallel between the training primitives behind Large Language Models and those of generative video models. > Curiously, the same primitives apply to today's generative video models. - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source and contributor context Google DeepMind · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Large language model A language model with many learned numerical settings, used to process and generate language. - [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model) ## Zero-shot Doing a task without worked examples (zero-shot) or with only a few examples given in the prompt (few-shot). - [Definition reference](https://developers.google.com/machine-learning/glossary#zero-shot-prompting) ## Foundation model A large model trained on broad data that can be adapted to many different tasks. - [Definition reference](https://developers.google.com/machine-learning/glossary#foundation-model) ## Google DeepMind Google DeepMind develops AI systems for general tasks and scientific research, including AlphaFold. - [Explore this context](/entities/google-deepmind) - [Reference](https://deepmind.google/blog/) - [Reference](https://deepmind.google/science/alphafold/) ## Editorial provenance Version 1. Publication ID: 589bd553-c386-c1fa-770c-394dd515022f. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: passed by anthropic/claude-opus-5.5 on 2026-11-12T03:38:58.085Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: c9657df5c0a812b993b4e224a1ec287ced3d108eec4bdecf80b026021434ce09. Source text hash: adcf9c9a800cfd0f92a2b45bebefff11034b4f047379c1f716e550de291ce348. ## Revision and source-check history ### Version 1 — 2026-11-12T03:43:26.106Z The paper was published on September 24, 2025. Thaddäus Wiedemer and coauthors report that Veo 3 can perform tasks it was not explicitly trained for, including segmenting objects, detecting edges, editing images and understanding physical properties. They say these abilities enable early visual reasoning, such as maze and symmetry solving. The authors draw a parallel between the simple primitives behind Large Language Models’ general-purpose language abilities—large generative models trained on web-scale data—and those used by generative video models. They argue that Veo’s reported zero-shot capabilities indicate video models are on a path toward unified, generalist vision foundation models. The paper’s argument matters because it presents video models as potentially able to understand and manipulate the visual world across tasks, not only generate video. The abstract does not detail how the capabilities were evaluated or establish how well they work beyond the reported examples. ## Original source Google DeepMind · Article Original source date: Sep 24, 2025 (2025-09-24T00:00:00+00:00). Date basis: Date stated on the publisher’s page. Added to Escape Meridian: 2026-11-12T03:43:26.106Z. First observed: 2026-10-29T01:09:48.146Z. Retrieved: 2026-10-29T01:11:40.256Z. - [Video models are zero-shot learners and reasoners](https://deepmind.google/research/publications/203190/) ## Limits The abstract does not detail how the reported capabilities were evaluated or how well they work beyond the listed tasks. Text as retrieved on October 29, 2026; originally published September 24, 2025. ## Open question What evidence would show whether the reported task abilities amount to general-purpose vision understanding? ## Source-stated Wiedemer and coauthors report that Veo 3 performs a broad range of tasks it was not explicitly trained for, including visual perception and editing tasks. > We demonstrate that Veo 3 can solve a broad variety of tasks it wasn't explicitly trained for: segmenting objects, detecting edges, editing images, understanding physical properties - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source-stated The authors say these abilities enable early forms of visual reasoning, including solving mazes and symmetries. > These abilities to perceive, model, and manipulate the visual world enable early forms of visual reasoning like maze and symmetry solving. - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source-stated The authors argue that Veo's emergent zero-shot capabilities indicate that video models are on a path toward unified, generalist vision foundation models. > Veo's emergent zero-shot capabilities indicate that video models are on a path to becoming unified, generalist vision foundation models. - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source-stated The authors draw a parallel between the training primitives behind Large Language Models and those of generative video models. > Curiously, the same primitives apply to today's generative video models. - [Supporting source](https://deepmind.google/research/publications/203190/) ## Source and contributor context Google DeepMind · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Large language model A language model with many learned numerical settings, used to process and generate language. - [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model) ## Zero-shot Doing a task without worked examples (zero-shot) or with only a few examples given in the prompt (few-shot). - [Definition reference](https://developers.google.com/machine-learning/glossary#zero-shot-prompting) ## Foundation model A large model trained on broad data that can be adapted to many different tasks. - [Definition reference](https://developers.google.com/machine-learning/glossary#foundation-model) ## Google DeepMind Google DeepMind develops AI systems for general tasks and scientific research, including AlphaFold. - [Explore this context](/entities/google-deepmind) - [Reference](https://deepmind.google/blog/) - [Reference](https://deepmind.google/science/alphafold/) ## Editorial provenance Version 1. Publication ID: 589bd553-c386-c1fa-770c-394dd515022f. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: passed by anthropic/claude-opus-5.5 on 2026-11-12T03:38:58.085Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: c9657df5c0a812b993b4e224a1ec287ced3d108eec4bdecf80b026021434ce09. Source text hash: adcf9c9a800cfd0f92a2b45bebefff11034b4f047379c1f716e550de291ce348. ### passed — 2026-11-12T03:38:58.085Z — anthropic/claude-opus-5.5 Publication: 589bd553-c386-c1fa-770c-394dd515022f attribution — pass: The reading attributes the claims to Thaddäus Wiedemer and coauthors in a Google DeepMind paper. This matches the listed authors and the publisher page. The text is a single paper abstract, not a listing. > Thaddäus Wiedemer, Effie Li, Paul Vicol, Shane Gu, Nick Matarese, Kevin Swersky, Been Kim, Priyank Jaini, Robert Geirhos summary — pass: The summary faithfully conveys the abstract: Veo 3 performs untrained tasks, early visual reasoning such as mazes and symmetry, the parallel with LLMs, and the claimed path to generalist vision foundation models. Zero-shot, LLM and foundation model are defined beside the reading. Veo 3 is identifiable from context as a generative video model. The title's 'reason' matches the paper's own 'reasoners' framing. The question is substantive and non-leading, and the stated limitation is appropriate. > Could video models be on a trajectory towards general-purpose vision understanding, much like LLMs developed general-purpose language understanding? date — pass: The reading states September 24, 2025, which matches the source dateline and evidence.originalPublished at day precision. It makes no post-publication or hindsight claims. The page was modified in December 2025, but the abstract is plainly the original paper text. The page kind is publisher-page, so no archive disclosure is needed. > September 24, 2025 support — pass: Each claim is directly stated in the abstract and kept as the authors' assertion. The significance sentence ('potentially able to understand and manipulate the visual world across tasks') is hedged and grounded in the source. The reading notes that evaluation details are absent rather than overstating reliability. The cited abstract is the core original text, so the later page modification does not cast doubt on it. > These abilities to perceive, model, and manipulate the visual world enable early forms of visual reasoning like maze and symmetry solving. media — pass: No media was selected and no verified media candidates were supplied, so no media identity or relevance issue arises. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/5ea5460e0bb5b641ae1eb9ea5406248470f1a4756018dc7eaf978dccdd84ea4c/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/thaddaus-wiedemer-and-coauthors-argue-in-google-deepmind-s-5ea5460e.json)