MiniMax H3 text-to-video demo with LTX 2.3 upscale draws praise as new king
alisitskii · reddit · 2026-08-04
A Reddit post showcases a “Mac and Cheese” clip made with MiniMax H3 text-to-video plus LTX 2.3 spatial upscaling.
The author says MiniMax H3 is now “the new king,” while also noting that image-to-video looks messier than text-to-video in this example. The attached frame shows a surreal cinematic shot, underscoring the model’s visual punch.
More from Fun
- Claude keeps pushing the annoying fix before five other ablations — pixlpa · 2026-08-04
- Sam Altman’s ChatGPT parenting idea gets a perfect one-line roast — xiaosun86 · 2026-08-04
- A new joke benchmark says AGI must catch a ball, hopscotch, and improvise street rhymes — Liu_eroteme · 2026-08-04
- Joke post riffs on universal-weave, a promptable DAG and CRDT loom library — amplifiedamp · 2026-08-04
- Old AI debate gets retold as a meme about rewriting history — AccBalanced · 2026-08-04
- LinkedIn’s AI slop now comes with AI-generated images of AI slop — SuB8u · 2026-08-04