MiniMax releases H3-Image
MiniMax's 11 November post releases open weights for an H3 image model that it says can make a 4096×4096 image in under 10 seconds.
Official source (opens in a new tab)
MiniMax's research post on 11 November 2026 releases MiniMax H3-Image, an open-weight image model built on H3. It covers text-to-image generation, reference-based generation, and image editing. MiniMax says it can generate a 4K image, 4096 by 4096, in under 10 seconds, at up to 5 times the speed of mainstream alternatives, and that it accepts up to 16 reference images. MiniMax also says prompt adherence is state of the art among open-weight models. Those speed and rank claims are MiniMax's.
The weights are out, in the same way MiniMax released H3. The model is also on the MiniMax API, MiniMax Code, and Hailuo AI. MiniMax says Standard mode is 1.5 to 2 times as fast as mainstream alternatives, and Highspeed mode is 3 to 4 times as fast. It prices the model at 50 to 70 percent of comparable mainstream offerings. For launch week, 4K and above is half the standard price.
Who it's for: someone generating or editing a still image from the weights, the API, or Hailuo, including a shot that has to follow several reference images.
Fits: Best for Images stays Midjourney for a considered frame, and Flux for prompt adherence among the widely used image models on that list. H3-Image is a new open checkpoint. MiniMax's own comparison does not move those two jobs. The image guide and the Flux alternatives are unchanged. The verdict is unchanged.