SenseTime open-sources SenseNova 6.8 Code2Video, a code-first editable video generation model
s_mohinii · x · 2026-11-05
SenseTime has released SenseNova 6.8 Code2Video as an open-source model that takes a code-first approach to video generation, emphasizing controllability over raw generation.
Key features:
- Generate animations directly from creative prompts
- Rebuild structure and motion of reference videos
- Refine creations conversationally without rebuilding from scratch
- Open weights, source code, creation Skills, and generation traces
Available on GitHub and Hugging Face, targeting creative coding, automated video production, and interactive visual experiences.
More from Multimodal
- Extract the last frame in Adobe Firefly Boards to keep AI shots consistent — Mr_AllenT · 2026-11-05
- Gaussian splat demo: 1.18 GB PLY compressed to 76 MB SOG with burnable scene — willeastcott · 2026-11-05
- splat-walk open-sourced: compressing a 1.18GB PLY to 76MB SOG for a fake phone-footage walkthrough — willeastcott · 2026-11-05
- Gaussian splat demo renders a photoreal 'phone walkthrough' of London in real time with PlayCanvas — willeastcott · 2026-11-05
- Building a $2,000 CAD ComfyUI PC: used RTX 3090 24GB vs newer GPUs? — Good_Business7425 · 2026-11-05
- Qwen 2.1 edit passes as real photo — and JPEG artifacts fool AI detectors — RadioQuick6509 · 2026-11-05