MiniMax H3 prompting for fight choreography: users still decoding what the model wants

krigeta1 · reddit · 2026-09-05

A Reddit thread discusses prompting best practices for MiniMax H3 video generation, focused on complex action and fight scenes.

• The author's workflow: use GPT-5.6 Sol with the official Ref2VA prompting guide to turn timing, characters, actions, camera moves, and environment into structured prompts.
• Pain point: formatting is easy, but getting H3 to understand exact choreography, positioning, motion direction, timing, and continuity often takes 5-6 prompt rewrites even for simple sequences.
• Open questions: literal vs. terse motion descriptions, second-by-second choreography, separating camera instructions from character actions, and preventing action swapping in two-character physical interactions.
• The thread aims to collect working prompt examples for fights, anime action, and live-action choreography as a community guide.

Original post →

More from Multimodal

• Voxtral 3 ships a LiveKit-based voice agent SDK streaming listen-think-speak — MistralAI · 2026-10-21
• Voxtral Mini TTS 3 hits SOTA expressiveness in 35+ languages with tag-based control — MistralAI · 2026-10-21
• Mistral launches Voxtral 3 speech models and new Voice Agents API — MistralAI · 2026-10-21
• Open-sourced pixel art normals generator ships as Rust CLI and wasm library — HugoDuprez · 2026-10-20
• Fibo 2.0 is strongest for ecommerce and creator content, weakest at UIUX design — ArtificialAnlys · 2026-10-20
• Fibo 2.0 leads old Fibo on all 9 capabilities, best at reasoning and anatomy — ArtificialAnlys · 2026-10-20

Multimodal channel →