Best image models to run locally on a 32GB Mac
A leaderboard of the best image models you can run locally on a Mac with 32GB of memory. Qwen-Image-2.1 is #1 right now, followed by Ideogram 4.0 and ERNIE-Image-Turbo. Models are ranked by one score that combines 73 published benchmarks. Each model shows the download that fits and the command to run it.
35 models · scores combine 73 benchmarks · updated how it works · find a leaderboard · benchmark your machine
Models for 32GB RAM · Apple Silicon → ~21GB fast · local only · reset
+ more
$ terminal
↗ share
detecting your OS…
Score vs memory to run
x
show
● open weights ◆ API only ─┘ best score for the memory · x: GB of RAM / VRAM for the smallest download (log)
| # | model | score↓ | where↓ | speed↓ | sources |
|---|---|---|---|---|---|
| 18 |  Qwen-Image-2.1 · int8 convrot + Qwen3-VL 8B q4_k_m (sd.cpp) | 69±2 | local |   | |
| 23 |  Ideogram 4.0 · mlx-q4 | 66±2* | local |  | |
| 47 |  ERNIE-Image-Turbo · gguf-ud-q4_k_m + Ministral 3B (sd.cpp) | 61±3† | local | | |
| 49 | ERNIE-Image · gguf-ud-q4_k_m + Ministral 3B (sd.cpp, all files in memory) | 60±2 | local |  | |
| 56 |  Krea 2 Turbo · gguf-q8_0 + Qwen3-VL 4B q4_k_m (sd.cpp) | 59±3* | local | | |
| 61 |  FLUX.2 [dev] · mlx-int4 (Flux2CLI, Swift) | 58±1 | local | +8 | |
| 63 |  Z-Image (Base) · gguf-q4_k_m + Qwen3 4B q4_k_m (sd.cpp) | 58±2 | local | | |
| 78 | Z-Image-Turbo · gguf-q3_k (+ Qwen3-4B Q4_K_M text encoder) | 55±1 | local | +1 | |
| 82 | FLUX.2 [klein] 9B · gguf-q4_0 + Qwen3 8B q4_k_m (sd.cpp) | 55±2 | local | | |
| 88 |  LongCat-Image · gguf-q4_k_m + Qwen2.5-VL q8_0 (sd.cpp) | 53±4 | local | | |
| 97 | FLUX.2 [klein] 4B · bf16 + Qwen3 4B bf16 (sd.cpp) | 52±2 | local | | |
| 113 |  Ovis-Image-7B · gguf-q4_0 + Ovis 2.5 (sd.cpp) | 48±7* | local | | |
| 126 | FLUX.1 Kontext [dev] · gguf-q4_k_m (sd.cpp) | 44±1 | local | +7 | |
| 127 | FLUX.1 [dev] · gguf-q4_0 | 44±1 | local |  +5 | |
| 135 | FLUX.1 [schnell] · mlx-4bit | 42±2 | local | | |
| 137 |  Stable Diffusion 3.5 Large · q8_0 (sd.cpp --type) | 42±1 | local | +4 | |
| 139 |  Playground v2 (1024px aesthetic) · fp16 (sd.cpp) | 41±4* | local | | |
| 148 | Stable Diffusion XL 0.9 (base) · q8_0 (sd.cpp --type) | 40±6* | local | | |
| 149 | Stable Diffusion 3 Medium · fp16 incl. CLIPs + T5 (sd.cpp) | 40±2 | local | | |
| 156 |  SDXL-Lightning (4-step UNet) · fp16 (sd.cpp) | 37±3* | local | | |
| 161 |  SSD-1B · fp16 (sd.cpp) | 36±4* | local | | |
| 167 | Stable Diffusion XL 1.0 · fp16 (sd.cpp) | 34±1 | local | +8 | |
| 170 | SDXL-Turbo · fp16 (sd.cpp) | 34±2 | local | | |
| 178 |  Dreamlike Photoreal 2.0 · q8_0 (sd.cpp --type) | 30±3 | local | | |
| 183 |  DALL·E Mega (minDALL-E) · fp32 | 29±4 | local | | |
| 185 |  Openjourney · q8_0 (sd.cpp --type) | 28±2 | local | | |
| 186 |  Stable Diffusion v1.4 · fp32 | 27±2 | local | | |
| 187 | Stable Diffusion v1.3 · fp32 | 27±4* | local | | |
| 188 | Stable Diffusion v1.1 · fp32 | 26±3 | local | | |
| 189 | Stable Diffusion v1.2 · fp32 | 26±3* | local | | |
| 191 | Stable Diffusion v1.5 · fp32 | 26±2 | local | +1 | |
| 193 |  LCM Dreamshaper v7 · openvino-fp16 (OpenVINO GenAI) | 23±4* | local | | |
| — |  Mage-Flow · bf16 + Qwen3-VL 4B q4_k_m (sd.cpp) | not yet calibrated | local | | |
| — | Mage-Flow-Base · bf16 + Qwen3-VL 4B q4_k_m (sd.cpp) | not yet calibrated | local | | |
| — | Mage-Flow-Turbo · bf16 + Qwen3-VL 4B q4_k_m (sd.cpp) | not yet calibrated | local | |
+ 167 more don't fit your machine or budget · show them
202 of 202 catalogue models listed here · sources and licences
* new or thinly tested: ranked on too few benchmark results · † only the maker’s own numbers so far · ~RAM is estimated; full context may need more memory