Images models that run on 36 GB

117 models with published weights that fit in 36 GB — a MacBook Pro with 36. Counted at four-bit quantisation, weights only, leaving the machine about a third of its memory and a gigabyte for context: room for roughly 37 billion parameters. A mixture of experts is counted in full, because it is held in full even though only a few experts compute. 117 in all, a hundred to a page; this is page 2 of 2.

An Apple configuration, and a comfortable one: it holds what 32 GB holds without the machine feeling tight, which in practice means a longer context or a browser you do not have to close first.

Models that make pictures from a description, and increasingly edit the ones you give them. Priced per image, sometimes by resolution, so the rate card reads nothing like a language model's. Beyond the look of the output, the practical differences are whether it can render legible text, whether it will edit rather than regenerate, and what the licence says about selling what comes out.

• Cyberpunk-Anime-DiffusionDGSpitzer0.9B≈0.6 GB at 4-bit
• Dungeons-and-Diffusion0xJustin0.9B≈0.6 GB at 4-bit
• EimisAnimeDiffusion_1.0veimiss0.9B≈0.6 GB at 4-bit
• Ghibli-Diffusionnitrosocke0.9B≈0.6 GB at 4-bit
• GuoFeng3xiaolxl0.9B≈0.6 GB at 4-bit
• Nitro-Diffusionnitrosocke0.9B≈0.6 GB at 4-bit
• Stable_Diffusion_PaperCut_ModelFictiverse0.9B≈0.6 GB at 4-bit
• anything-v5Stable Diffusion API0.9B≈0.6 GB at 4-bit
• dreamlike-anime-1.0dreamlike-art0.9B≈0.6 GB at 4-bit
• dreamlike-diffusion-1.0dreamlike-art0.9B≈0.6 GB at 4-bit
• dreamlike-photoreal-2.0dreamlike-art0.9B≈0.6 GB at 4-bit
• openjourney-v4prompthero0.9B≈0.6 GB at 4-bit
• PixArt-XL-2-1024-MSPixArt0.6B≈0.4 GB at 4-bit
• control_v1p_sd15_brightnessLatent Cat0.4B≈0.2 GB at 4-bit
• controlnet_qrcode-control_v1p_sd15DionTimmer0.4B≈0.2 GB at 4-bit
• FLUX.1-Turbo-Alphaalimama-creative0.3B≈0.2 GB at 4-bit
• openjourneyprompthero0.1B≈0.1 GB at 4-bit

Wider