Verified, then logged.
Sun 22 Nov 2026 · 00:15 KST / 15:15 UTC한국어 | ENSearch /

01 / AI · Claim vs. Check · Analysis

MiniMax unveils model for direct high-resolution images

DigitalToday reported that MiniMax's image model supports generation without separate upscaling, editing and distributed weights for developers to run it themselves.

Through the windshield of a small aircraft on approach at overcast dawn, a runway's white straight lines converge toward the horizon, its far end hidden by fog and one side cut off by the window frame. AI-generated image

MiniMax, a Chinese AI company, unveiled a model that generates images of sixteen million pixels without separate upscaling, according to DigitalToday, a Korean technology news publication.

The report described both the image size and the generation process. The model produces high-resolution images directly, removing the separate enlargement step described as upscaling.

DigitalToday cited Japanese publication Gigazine in its coverage. Its report also covered image editing and the distribution of model weights for developers to run the model in their own environments.

Image generation and editing

DigitalToday identified the model as MiniMax-H3-Image. It reported that the image model was developed from MiniMax H3, the company's video-generation AI.

According to the report, the model supports image editing and the use of multiple images as references. It also offers faster high-resolution image generation.

Model weights for developers

DigitalToday reported that MiniMax also distributed the model weights, the learned parameters used to run a model. The distribution allows developers to run it in their own environments.

The release therefore covers both image capabilities and access for developers, according to the report. Alongside generating and editing images, developers can run the model themselves using the distributed weights.