MiniMax H3
MiniMax's text-to-video and image-to-video model, and the base that studio-1939-lora adapts. What
the corpus holds is the shape of it rather than the model itself: LoRA adapters are supported and
served — fal exposes minimax/h3/text-to-video/lora as a standard endpoint — and the weights carry a
MiniMax H3 Community License rather than Apache-2.0 or MIT, so it sits in the same
open-but-conditional band this spoke records for several Chinese releases.
Thin page, and deliberately so. No parameter count, architecture, benchmark, price or release date was read at ingest; the only source here is an adapter's model card pointing at it. Anything about the model's own capability or cost needs its own source.
Related
studio-1939-lora · llm-provider · open-weight-models · meridian