Will It Run AI · Diffusion Calculator
Check if your GPU can run image and video generation models locally.
Pick your hardware, choose a diffusion model, and see the VRAM breakdown, generation time, and maximum resolution in real time.
Configuration
Default: 28
Results
C
Runs with offload
Your RTX 4090 24GB can run Flux.1 Dev at 1024 x 1024
VRAM Breakdown
Required: 24.0 GBAvailable: 24.0 GB
Weights24.0 GB
VAE0.2 GB
Text Encoder9.6 GB
Activations0.8 GB
Overhead0.5 GB
Per Image
~18s
Max Resolution
256×256
VRAM Used
24 / 24GBMeasured
Flux.1 Dev
12B parameters · DIT · 28 steps
State-of-the-art text-to-image model from Black Forest Labs. Excels at photorealism, text rendering, and prompt adherence. 12B parameter DiT architecture with dual text encoders: T5-XXL (4.7B) and CLIP-L (0.12B).
Speed & VRAM Tips
FP8 available: ~1.5x faster, ~45% less VRAM on RTX 40xx GPUs. Try switching precision to FP8.
Turbo/LCM: Reduce steps to 4-8 with a distilled scheduler for 7x faster generation (slight quality trade-off).
GGUF Q4: Flux has GGUF quantized versions (6.8GB vs 23.8GB). Use NF4 precision for a similar effect.