cd /news/generative-ai/pruna-ai-shrinks-minimax-h3-from-143… · home › topics › generative-ai › article
[ARTICLE · art-172623] src=alphasignal.ai ↗ pub= topic=generative-ai verified=true sentiment=↑ positive

Pruna AI Shrinks MiniMax-H3 From 143 GB Down to a Single GPU

Pruna AI released P-Image-OSS, an open-weights image generation and editing model rebuilt from MiniMax's 143 GB MiniMax-H3 video model, cutting the smallest package to a 20 GB download that peaks at about 21 GB of VRAM and fits a 24 GB graphics card. The model generates 8-megapixel images in 8 denoising steps without classifier-free guidance, roughly 5x faster than H3's 50 steps, and has recorded more than 1.2 million downloads through its ComfyUI repository on Hugging Face. Pruna distilled H3's 32-billion-parameter text encoder to 4 billion parameters (67 GB to 9 GB), cut the transformer to 20B and trained a new decoder, though the MiniMax license excludes the EU, UK, South Korea and US without a separate MiniMax license.

key takeaway

P-Image-OSS runs 8 MP generation in 8 steps without CFG at about 21 GB peak VRAM, versus 143 GB of weights for MiniMax-H3, by distilling the text encoder from 32B to 4B parameters (67 GB to 9 GB) and cutting the transformer to 20B.

read3 min views2 publishedNov 19, 2026
Pruna AI Shrinks MiniMax-H3 From 143 GB Down to a Single GPU
Image: Alphasignal (auto-discovered)

Pruna AI turned MiniMax's 143 GB H3 video model into an open image generator and editor that renders 8 MP images in 8 steps and fits in 21 GB of VRAM.

  • Pruna AI released P-Image-OSS , an open-weights image model rebuilt from the MiniMax-H3 video model.
  • It makes 8 MP images in 8 steps without CFG, about 5x faster than H3's 50 steps.
  • The smallest 4-bit set is a 20 GB download peaking at 21 GB VRAM, versus 143 GB for H3.
  • Pruna distilled the text encoder from 32B to 4B, cut the transformer to 20B, and trained a new decoder.
  • It edits with up to 8 reference images and runs natively in ComfyUI and Diffusers .
  • The MiniMax license excludes the EU, UK, South Korea and US without a separate MiniMax license.

Pruna AI converted MiniMax-H3, a video model with 143 GB of weights, into an image generation and editing system. The resulting P-Image-OSS, initially announced as P-H3-Image, has recorded more than 1.2 million downloads through its ComfyUI repository on Hugging Face.

P-Image-OSS generates 8-megapixel images in eight denoising steps, the iterative passes that turn noise into an image. It runs without classifier-free guidance, a common technique that evaluates conditioned and unconditioned predictions at each step and roughly doubles denoiser work. The model can also edit with up to eight reference images. Its smallest package is a 20 GB download with a reported peak of about 21 GB of VRAM, placing it within the memory budget of a 24 GB graphics card.

Why a video backbone helps #

Video models must preserve objects, lighting, materials and anatomy as scenes change across frames. That training gives them useful representations of three-dimensional shape, motion and viewpoint, which can transfer to still-image generation. Pruna previously tested the approach with Wan Image, built from Wan 2.1, before applying it to MiniMax-H3 after MiniMax released the weights.

H3’s original design created two obstacles for image generation: individual frames appeared soft, and the model required too much memory and compute for common local hardware. Pruna’s technical report describes changes to the prompt upsampler, text encoder, diffusion schedule, transformer and decoder.

Five changes behind the conversion #

Prompt upsampler. H3 responds strongly to prompt structure. Pruna trained a Qwen3.5-9B-based model to expand a short request into the detailed format H3 handles well. It processes the request before drafting the expanded prompt, uses greedy decoding for repeatable output and limits prompt length to reduce glossy or overworked images. 2. Text encoder. H3 originally interpreted prompts with a 32-billion-parameter encoder. Pruna distilled that encoder into a 4-billion-parameter model by training the smaller network to reproduce the larger model’s representations. The reported footprint fell from 67 GB to 9 GB.

This story is for Pro members #

You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.

── more in #generative-ai 4 stories · sorted by recency
── more on @pruna ai 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/pruna-ai-shrinks-min…] indexed:0 read:3min 2026-11-19 · —