According to the current roadmap, improvements are expected in:

visual quality audio motion dynamics stylization consistency Planned new capabilities include:

first-frame / last-frame guided video generation 9-grid image-to-video subject and voice reference instruction-based video editing video recreation / replication The overall direction seems to be a more complete video generation and editing workflow rather than a single-feature update.

Atlascloud.ai has been in contact with the Wan team and plans to support Wan 2.7 on launch day.

Would be interesting to see benchmark examples and workflow demos once it ships.