C-RADIOv4 (Tech Report)
Mike Ranzinger ∗ Greg Heinrich ∗ Collin McCarthy Jan Kautz Andrew Tao Bryan Catanzaro Pavlo Molchanov · 2026
Read C-RADIOv4 (Tech Report) by Mike Ranzinger ∗ Greg Heinrich ∗ Collin McCarthy Jan Kautz Andrew Tao Bryan Catanzaro Pavlo Molchanov (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-17237.txt.
Description
Mike Ranzinger and six co-authors report on C-RADIOv4, a vision model for image tasks. The model learns from many teacher models at once. That method is called multi-teacher distillation. A single student model keeps the skills of each teacher. It can even improve on them in places. The design builds on the AM-RADIO and RADIOv2.5 models before it.
The report says the new model scores better on key tasks that come after training. It does so at the same compute cost as before. Such shared backbones serve many vision jobs at once. Teams can use one base for search, tagging, and more. The report tracks what changed in this release.
Catalog record
| CREATOR | Mike Ranzinger ∗ Greg Heinrich ∗ Collin McCarthy Jan Kautz Andrew Tao Bryan Catanzaro Pavlo Molchanov |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-17237