H3-Tiny-17B-Pruned-FL2VA vs PAN 2
Compare documented capabilities and access. A dash means a comparable figure is unavailable.
| Specification | ||
|---|---|---|
| Task | Audiovisual generation | World simulation |
| Published capabilities | Generates synchronized video and audio from a prompt or from first and last frames. | Predicts what happens after an action in a starting scene. |
| Model size | 17.07B transformer parameters | — |
| Access | Downloadable transformer checkpoint for the H3 pipeline | Public weights, full training code and inference code |
| Result | — | Simulated video conditioned on a scene and action |
| Operating limits | 50 steps | — |
| Sources | Model weights | Publisher documentation |
| Inputs | Text, Images | Text, Images |
| Weights | Open weights · MiniMax H3 Community License | Open weights · Apache-2.0 |
| Released | Nov 20, 2026 | Nov 10, 2026 |