Testing Minimax H3: better results with 32+ steps
rm_rf_all_files · reddit · 2026-08-17
A user shared their second take on generating videos with Minimax H3. The first attempt was a low-res preview (0.3MP) which suffered from morphing due to fast motion. This test used 32+ steps, and the user provided a link to the result, indicating improved performance at higher step counts.
More from Multimodal
- EgoTools: 100-hour egocentric video dataset teaches AI tool-centric reasoning — liuziwei7 · 2026-10-03
- Redditor Gets YEDP UV Painter Working with Flux Klein, Shares the Workflow — o0ANARKY0o · 2026-10-03
- PixAl Releases Tsubaki.3 Anime Model, Publishes Report on Style Diversity — Level-Ninja-2492 · 2026-10-03
- How AI talking-head channels pump out daily videos: four lip-sync tools tested and the cost problem — No_Shoe1628 · 2026-10-03
- Runway AI Summit closes with Valenzuela reflection, Labs unveils Continuum — runwayml · 2026-10-03
- Experimenting with AI outpainting to revive and extend old photos — rufusd · 2026-10-03