Fal released h3 max, a video model built on open-weights minimax h3 that generates five seconds of…
fal released h3 max, a video model built on open-weights minimax h3 that generates five seconds of footage in about three seconds of wall time, roughly 35 times the throughput of the official endpoint. the open-weights video race is being won on serving economics, not on prompt quality.
a model you can run cheap beats a better model you cannot.
Context
fal's blog of 27 August 2026 says H3 Max is a post-trained version of MiniMax H3 developed by fal Research, starting from the open-weights MiniMax H3 model with added post-training data and a co-designed inference stack, available on fal's platform. It claims a 5-second video in approximately 3 seconds and roughly 35x the throughput of the official MiniMax H3 endpoint. It says fal's human-preference evaluation ranks it first among twelve video models and that Artificial Analysis and Design Arena also rank it first.
These are vendor-reported figures, not independent, and the Artificial Analysis and Design Arena rankings were not inspected. Hardware, resolution and batch conditions for the speed figures were not read. The 35x is against MiniMax's official endpoint, a service comparison, not a model-only comparison. The post supports the base being open-weights, and whether H3 Max's own weights are open was not stated in the text read. The release was 27 August 2026, three weeks before the note. The open-weights video race is being won on serving economics is the author's conclusion, and the post also claims better quality.
Related work
- H3 Max built with fal inference and training (fal, 17 Sep 2026) ↗Snippet only.
- H3 Max text-to-video (fal model page) ↗Snippet only.
Watch next
- Independent latency and quality tests, and whether H3 Max weights are released.
Sources
- Introducing H3 Max by fal (fal, 27 Aug 2026)blog.fal.ai
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 18 September 2026 at 20:49 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →