MiniMax H3 ref2v generates classic-anime Genshin clips — inconsistent style but surprising results
ForsakenContract1135 · reddit · 2026-08-29
The author used MiniMax H3's ref2v (image-to-video) to generate classic-anime-style Genshin Impact clips. Results exceeded expectations, though the art style is clearly inconsistent — the reference images themselves had different styles.
Workflow details:
- Mostly the default workflow, adding seg attention and a light2x turbo LoRA (8 steps), sampler euler+beta.
- Clips stitched at 8–10 seconds each, 0.7mp.
- Prompts generated with Grok's Minimax H3 skill: upload the image, say it's ref2v, roughly describe the action — any LLM works; GPT gives better results overall but the author prefers Grok.
Related event: Hands-On: MiniMax H3 ref2v Generates Genshin-Style Anime Videos(2 posts)→
More from Multimodal
- HyperFrames launches Reference Any Video and auto-applied Brand Kits for AI video — aftahi_ai · 2026-10-16
- Krea 2 + LoRA still beats Flux 3 Klein in pixel-art style test — Angaisb_ · 2026-10-16
- World Labs starts rolling out Atlas beta with pixel-perfect camera path control — theworldlabs · 2026-10-16
- Krea 3 image editing model lands on fal: high-quality edits in under 6 seconds — gorkem · 2026-10-16
- Tripo P2 vs Meshy 7.1: benchmarking today's best image-to-3D models for game assets — noahsolomon · 2026-10-16
- BytePlus-backed AI Film Award offers $1M grand prize for films 70% made with Seedance — DavidmComfort · 2026-10-16