Z-Image-Turbo (6B S3-DiT) VRAM & GPU Requirement Calculator
Interactive hardware sizing for Z-Image-Turbo (6B S3-DiT) (6B by Tongyi MAI). Estimated 8K context footprint: 4.8 GB (INT4) • 7.3 GB (FP8) • 13.3 GB (FP16).
GPUs
Memory Requirements 13.34 GB
Requires 1 GPUs (based on memory capacity)
12 GB
All model weights
0 GB
Conversation history cache
0.62 GB
Concurrent processing cache
0.72 GB
Temporary computation cache
ComfyUI Pipeline & Hardware Estimates
Est. Gen Time
2.5 s/img
Min System RAM
32 GB RAM
Resolution
1 MP
Offload Mode
Smart Offload
Quick Deploy CLI Command
python main.py --normalvram --preview-method auto
Flags reflect your selected precision, context length, and tensor parallel GPU count.Open Z-Image-Turbo spec page →
Z-Image-Turbo (6B S3-DiT) Architecture & Memory Profile
Total Parameters
6B
Architecture
Dense Transformer
Developer / Provider
Tongyi MAI
Recommended Precision
FP16
Z-Image-Turbo (6B S3-DiT) ComfyUI VRAM Matrix by Precision & Resolution
Precomputed 3-stage ComfyUI Smart Offload + VAE Tiling VRAM peak (recommended host RAM: 32 GB).
| Precision | 1024x1024 (1MP) | 1536x1536 (2.25MP) | 2048x2048 (4MP) | LoRA Train (1MP) |
|---|---|---|---|---|
| FP16 | 13.3 GB | 14.1 GB | 15.6 GB | 16.5 GB |
| FP8 | 7.3 GB | 8.1 GB | 9.6 GB | 10.5 GB |
| INT8 | 7.3 GB | 8.1 GB | 9.6 GB | 10.5 GB |
| INT4 | 4.8 GB | 5.6 GB | 7.1 GB | 8.0 GB |
Recommended GPU Hardware for Z-Image-Turbo (6B S3-DiT) (FP16)
Minimum GPU count for 1024x1024 generation with ComfyUI offload and estimated seconds/image.
| GPU Hardware | Min GPUs | Total VRAM Pool | Est. Gen Time |
|---|---|---|---|
| NVIDIA RTX 4090 (24GB) | 1× | 24 GB | 2.5 s/img |
| NVIDIA RTX 5090 (32GB) | 1× | 32 GB | 2 s/img |
| Apple M4 Max (128GB) (128GB) | 1× | 128 GB | 3 s/img |
| NVIDIA A100 (80GB) (80GB) | 1× | 80 GB | 2.6 s/img |
| NVIDIA H100 (80GB) | 1× | 80 GB | 0.7 s/img |
| NVIDIA H200 (141GB) | 1× | 141 GB | 0.7 s/img |
| Huawei Ascend 910B (64GB) | 1× | 64 GB | 1.4 s/img |
Z-Image-Turbo (6B S3-DiT) Deployment FAQ
How much VRAM is needed to run Z-Image-Turbo (6B S3-DiT)?
At 1024x1024 with ComfyUI Smart Offload and VAE Tiling enabled, Z-Image-Turbo (6B S3-DiT) needs about 13.3 GB in BF16/FP16, 7.3 GB in FP8, and 4.8 GB in INT4/GGUF (paired with ~32 GB system RAM).
Can I run Z-Image-Turbo (6B S3-DiT) on RTX 4090 or H100 GPUs?
Running Z-Image-Turbo (6B S3-DiT) in FP16 precision requires at least 1x RTX 4090 (24GB) or 1x NVIDIA H100 (80GB) GPUs.
What precision should I choose for Z-Image-Turbo (6B S3-DiT) in production?
For Z-Image-Turbo (6B S3-DiT) (6B by Tongyi MAI), FP16 provides the best balance between inference speed, VRAM footprint, and output quality.