SenseNova open-sources Code2Video model that writes and revises code to make videos
Negative-Whereas3307 · reddit · 2026-10-27
SenseNova 6.8 Code2Video is an open-source release following the "model writes code, renderer produces video, natural language revises it" approach, focused on programmatically expressible visuals like motion graphics.
Key points:
- Generates video via code, targeting motion graphics and programmatic visuals
- Supports natural-language revisions of styling, timing, and scene elements
- Can reconstruct a reference video's visual treatment as editable code
- Paired with a creation Skill guiding briefs, previews, and iterations
- Full weights plus the Skill to be released; INT4/INT8 planned
The open question: whether a dedicated Code2Video model beats general coding models on fewer repair attempts, better motion, and safer targeted edits.
More from coding & agent
- Anthropic Reveals 12 New Prompting Rules for Claude Opus 5.5 — JeremyCMorgan · 2026-10-28
- TraceTrial-Bench separates agent success from harness reliability: 31.1% of runs mislead — _akhaliq · 2026-10-28
- NVIDIA open-sources a LoRA post-training workflow to localize Nemotron with synthetic personas — NVIDIAAI · 2026-10-28
- Hobbyist builds 'Sandman' bot fleet dashboard to gamify context management — alexcovo_eth · 2026-10-28
- Apple's iPrefBench: mobile agents execute well (91.6%) but personalize poorly (47.7%) — apple · 2026-10-28
- Alibaba Tongyi MAI releases MAI Gym, a 9B open GUI agent hitting SOTA under 10B — AdinaYakup · 2026-10-28