Seedance 2.5 Tested: Generating Coherent Video from Two Reference Images
chrisfirst · x · 2026-08-05
A user shared a hands-on test of the Seedance 2.5 video generation model. The demonstration uses two reference images along with a detailed prompt to generate a coherent video of a character walking through a market, sitting down, and being served by a robot, showcasing the model's capability in multi-image reference and complex action understanding.
More from Multimodal
- Qwen-Image-3.0 Released: Ranks #1 in China, Supports 4.5k-Token Prompts — arena · 2026-08-05
- Peking University Introduces UniWorld-Design for Layer-Native Image Generation — PekingUniversity · 2026-08-05
- JD's Any-OPD Enables Cross-Architecture Distillation for Flow-Matching Image Models — jingdong1 · 2026-08-05
- Runway H3 Model Confuses Character Voices in Multi-Character Dialogues — ptwonline · 2026-08-05
- Running 1.5B Voice Model Locally on iPhone: Only 2.2GB Memory — Acceptable-Cycle4645 · 2026-08-05
- MiniMax H3 Reported as the Most Uncensored AI Video Model Available — beechinour · 2026-08-05