SYLIN

MiniMax H3 Image

by MiniMax · Model · Image

strong

One of the best open image models, fast and close to Qwen Image 2.1

Open image model generating 4K images in seconds with natural-language editing and reference images; called one of the best open image generators

AccessOpen weightsNeeds~204 GB; smallest build ~6 GBDownloadabout 204 GB full repository; 12.5 GB for the W4A8 versionSize33B parametersLicenseApache 2.0

Why we say so

Artificial Analysis ranks it first among open-weights models for editing and text-to-image, and Arena puts it second for open editing (1358, just behind Qwen Image 2.1's 1363) and third for open text-to-image. AI Search found its edits very close to Qwen Image 2.1, but Krea 2 and MOSS-Image beat it on some realism prompts, and it does poorly at 1 megapixel. Fahd Mirza saw garbled small text and failures on complex scenes, and reported VRAM use from 16 GB (AI Search) to nearly 48 GB, against MiniMax's 6 GB claim.

Ranks #2 for image editing (open weights) · #3 for image generation (open weights)

Evidence

  1. Nov 17
    #2 on Arena image-edit among open weights (1358)
  2. Nov 17
    #3 on Arena text-to-image among open weights (1198)
  3. Nov 15
    Open-weight 33B diffusion transformer distilled to 6 sampling steps, no CFG, native ~16 MP output without upscaler
  4. Nov 14
    Reviewer calls it definitely among the best open image editors, close to Qwen Image 2.1
  5. Nov 12
    Open image generation and editing model released; runs locally in ComfyUI with day-one workflow templates
Show all 8
  1. Nov 12
    Edit quality very close to Qwen Image 2.1; better at fixing lighting and colour, less faithful when restoring old photos
  2. Nov 12
    Krea 2 better at realism, yoga pose anatomy, dot illustration and Monet style; MiniMax wins UI design, magazine cover, soles prompt
  3. Nov 12
    Reviewer thinks Krea 2 or MOSS-Image are still better for text-to-image in some respects, but MiniMax is far faster at 2K/4K

Cost to run

Alternatives

Qwen Image#1 on Arena image-edit among open weights (1363)
HunyuanImage Instruct#3 on Arena image-edit among open weights (1301)
Ideogram#2 on Arena text-to-image among open weights (1199)
Built on MiniMax H3 Image