FLUX.1-dev (12B DiT + 4.9B TE) VRAM & GPU Requirement Calculator

Interactive hardware sizing for FLUX.1-dev (12B DiT + 4.9B TE) (11.9B by Black Forest Labs). Estimated 8K context footprint: 8.4 GB (INT4) • 13.4 GB (FP8) • 25.4 GB (FP16).

GPUs

Memory Requirements 13.45 GB

Requires 1 GPUs (based on memory capacity)

11.9 GB

All model weights

0 GB

Conversation history cache

0.9 GB

Concurrent processing cache

0.65 GB

Temporary computation cache

ComfyUI Pipeline & Hardware Estimates
Est. Gen Time
7.5 s/img
Min System RAM
32 GB RAM
Resolution
1 MP
Offload Mode
Smart Offload
Quick Deploy CLI Command
python main.py --normalvram --fp8_e4m3fn-unet --fp8_e4m3fn-text-enc --preview-method auto
Flags reflect your selected precision, context length, and tensor parallel GPU count.Open FLUX.1-dev spec page →

FLUX.1-dev (12B DiT + 4.9B TE) Architecture & Memory Profile

Total Parameters
11.9B
Architecture
Dense Transformer
Developer / Provider
Black Forest Labs
Recommended Precision
FP8

FLUX.1-dev (12B DiT + 4.9B TE) ComfyUI VRAM Matrix by Precision & Resolution

Precomputed 3-stage ComfyUI Smart Offload + VAE Tiling VRAM peak (recommended host RAM: 32 GB).

Precision1024x1024 (1MP)1536x1536 (2.25MP)2048x2048 (4MP)LoRA Train (1MP)
FP1625.4 GB26.7 GB28.9 GB29.1 GB
FP813.4 GB14.8 GB17.0 GB17.2 GB
INT813.4 GB14.8 GB17.0 GB17.2 GB
INT48.4 GB9.8 GB12.0 GB12.2 GB

Recommended GPU Hardware for FLUX.1-dev (12B DiT + 4.9B TE) (FP8)

Minimum GPU count for 1024x1024 generation with ComfyUI offload and estimated seconds/image.

GPU HardwareMin GPUsTotal VRAM PoolEst. Gen Time
NVIDIA RTX 4090 (24GB)1×24 GB7.5 s/img
NVIDIA RTX 5090 (32GB)1×32 GB5.4 s/img
Apple M4 Max (128GB) (128GB)1×128 GB9 s/img
NVIDIA A100 (80GB) (80GB)1×80 GB7.9 s/img
NVIDIA H100 (80GB)1×80 GB2.2 s/img
NVIDIA H200 (141GB)1×141 GB2.1 s/img
Huawei Ascend 910B (64GB)1×64 GB4.2 s/img

FLUX.1-dev (12B DiT + 4.9B TE) Deployment FAQ

How much VRAM is needed to run FLUX.1-dev (12B DiT + 4.9B TE)?
At 1024x1024 with ComfyUI Smart Offload and VAE Tiling enabled, FLUX.1-dev (12B DiT + 4.9B TE) needs about 25.4 GB in BF16/FP16, 13.4 GB in FP8, and 8.4 GB in INT4/GGUF (paired with ~32 GB system RAM).
Can I run FLUX.1-dev (12B DiT + 4.9B TE) on RTX 4090 or H100 GPUs?
Running FLUX.1-dev (12B DiT + 4.9B TE) in FP8 precision requires at least 1x RTX 4090 (24GB) or 1x NVIDIA H100 (80GB) GPUs.
What precision should I choose for FLUX.1-dev (12B DiT + 4.9B TE) in production?
For FLUX.1-dev (12B DiT + 4.9B TE) (11.9B by Black Forest Labs), FP8 provides the best balance between inference speed, VRAM footprint, and output quality.