X · X// 4h agoMODEL RELEASE

DeepInfra adds Wan2.7 video models

DeepInfra is now offering two Wan 2.7 video models: Wan2.7-I2V converts still images into cinematic motion while keeping subjects, style, and text stable, and Wan2.7-R2V uses reference media to preserve character appearance and voice across scenes. Both models are priced at $0.10 per second, positioning them as accessible options for developers building image-to-video and reference-driven video workflows.

// ANALYSIS

Strong platform expansion for teams that care about consistency and continuity more than experimental novelty.

–Wan2.7-I2V is the cleaner fit for image-led generation where motion quality matters but fidelity to the source image must stay high.
–Wan2.7-R2V is the more compelling differentiator, since keeping look and voice aligned across scenes is hard and valuable for character-centric video.
–The pricing is simple and low-friction for testing, but production usage can still add up fast on longer clips.
–This looks like a practical infrastructure release rather than a flashy consumer launch.

// TAGS

deepinfrawan2.7wan2.7-i2vwan2.7-r2vai-videoimage-to-videoreference-to-videovideo-generationinference

DISCOVERED

4h ago

2026-04-29

PUBLISHED

5h ago

2026-04-29

RELEVANCE

8/ 10

AUTHOR

DeepInfra