{"version":1,"type":"story","url":"https://digestai.news/story/nvidia-researchers-introduce-physis-lang-boosting-cosmos3-past-veo-3-1","json":"https://digestai.news/story/nvidia-researchers-introduce-physis-lang-boosting-cosmos3-past-veo-3-1.json","markdown":"https://digestai.news/story/nvidia-researchers-introduce-physis-lang-boosting-cosmos3-past-veo-3-1.md","slug":"nvidia-researchers-introduce-physis-lang-boosting-cosmos3-past-veo-3-1","headline":"NVIDIA researchers introduce Physis-Lang, boosting Cosmos3 past Veo 3.1","summary":"A joint team from NVIDIA, MIT and the University of Oxford presented Physis-Lang, a framework that treats physical language as an optimizable representation for video models. The system adds a physics‑reasoning field to captions and uses a self‑evolving loop where a frozen GPT‑5.5 captioner is guided by a Gemini‑3.1‑Pro critic, improving caption F1 from 78.64 to 87.82 after nine iterations. On the Physics‑IQ Verified leaderboard snapshot of September 29, 2026, Physis‑Lang on Cosmos3‑Super achieved a top score of 48.2 ± 1.4, while the Cosmos3‑Nano version placed second at 43.3 ± 1.5.\n\nWhen compared with Google’s Veo 3.1, Cosmos3‑Nano equipped with Physis‑Lang outperformed on three of four benchmarks, for example scoring 43.41 versus 34.99 on Physics‑IQ Verified. The researchers also distilled the pipeline into two local Qwen‑based models, PhysThinker‑C and PhysThinker‑U, which retained most of the gains while cutting API costs from about $24.12K to $0.12K, and a fully local setup cost $0. No code or weights have been released yet; only the paper and repository are available.","keyPoints":["Physis-Lang framework adds physics‑reasoning captions, reaching 48.2 ± 1.4 on Physics‑IQ Verified with Cosmos3‑Super.","Cosmos3‑Nano with Physis‑Lang beats Google Veo 3.1 on three of four benchmarks, e.g., 43.41 vs 34.99 on Physics‑IQ Verified.","Local PhysThinker‑C/U models keep most gains, reducing API expense from $24.12K to $0.12K."],"whyItMatters":"Shows language‑driven self‑evolution can close physics gaps in video models without architectural changes, lowering compute cost and improving benchmark performance.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["NVIDIA","Google","MIT","University of Oxford"],"models":["Cosmos3-Super","Cosmos3-Nano","Veo 3.1","Physis-Lang","PhysThinker-C","PhysThinker-U"],"people":["Asif Razzaq"]},"firstPublishedAt":"2026-09-30T07:32:34Z","updatedAt":"2026-09-30T07:32:34Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"MarkTechPost","title":"NVIDIA Researchers Introduce Physis-Lang: Self-Evolving Physical Language That Lifts Cosmos 3 Past Veo 3.1 on Physics Benchmarks","url":"https://marktechpost.com/2026/09/30/nvidia-researchers-introduce-physis-lang-self-evolving-physical-language-that-lifts-cosmos-3-past-veo-3-1-on-physics-benchmarks","publishedAt":"2026-09-30T07:32:34Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"NVIDIA researchers introduce Physis-Lang, boosting Cosmos3 past Veo 3.1\", 30 September 2026, https://digestai.news/story/nvidia-researchers-introduce-physis-lang-boosting-cosmos3-past-veo-3-1","publisher":"Digest AI","title":"NVIDIA researchers introduce Physis-Lang, boosting Cosmos3 past Veo 3.1","datePublished":"2026-09-30T07:32:34Z","url":"https://digestai.news/story/nvidia-researchers-introduce-physis-lang-boosting-cosmos3-past-veo-3-1"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}