All models

Wan / VIDEO

Wan 2.1 T2V 1.3B VRAM requirements

Explore memory tradeoffs with sourced model data and transparent estimates.

FP8 estimate

12.8 GiB

at default settings

FP16 estimate

12.8 GiB

at default settings

Suggested GPU

1 × NVIDIA T4 16 GiB

Model files & facts

Mediavideo
Default resolution832 × 480
Default frames81
Temporal compression4
NotesFile sizes from the Comfy-Org ComfyUI repackage.
ComponentWeight file (source)Size
Backbone FP16wan2.1_t2v_1.3B_fp16.safetensors2.64 GiB
UMT5-XXL FP16umt5_xxl_fp16.safetensors10.59 GiB
UMT5-XXL FP8umt5_xxl_fp8_e4m3fn_scaled.safetensors6.27 GiB
VAEwan_2.1_vae.safetensors0.24 GiB

Precision & offloading comparison

Source default resolution and frame count · batch 1 · FP16 text encoders · safety margin included.

BackboneText encodersTotal
FP16GPU12.75 GiB
FP16CPU6.99 GiB

ComfyUI file placement

Example component locations; install matching loader nodes. Brace patterns in source filenames denote all shards. Preserve the source layout when a loader requires it.

models/diffusion_models/wan2.1_t2v_1.3B_fp16.safetensors
models/text_encoders/umt5_xxl_fp16.safetensors
models/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors
models/vae/wan_2.1_vae.safetensors

Recommended GPUs by memory fit

FP16 backbone at source default resolution, FP16 text encoders on GPU. Smallest nominal memory first; green results only. Check backend and precision support separately.

Sources & verification

Hugging Face model card

Exact file sources are linked in the component table above. ComfyUI memory planner constant · Measured cross-check

Last verified: 2026-11-18

Try your configuration

Your workload

Custom diffusion model parameters

Uses the ComfyUI activation and decode profile for Wan 2.1 T2V 1.3B. Leave fields empty to use catalog files. Placeholder counts approximate half-precision file sizes.

Your hardware

Output dimensions

Images or videos per batch. Serving concurrency and throughput require separate measurements.

Resolution and peak VRAM
Sampling and timing

One step is one scheduler iteration. Enter timings from the same device, batch, resolution and guidance settings. Fixed overhead includes encoding, decoding and other work outside sampling.

Advanced
Decode at full size
Off — text encoders stay on the GPU
Off — ComfyUI swaps models between stages; the peak stage counts

Performance and memory results

53.1%Memory use

Memory estimate

12.8 GiB needed

Fits11.2 GiB to spare24.0 GiB available

The estimated workload fits with the safety margin included.

Wan 2.1 T2V 1.3B

RTX 4090 24 GiB

Model specifications
Memory breakdown · text encode stage (peak)GiB

Vertical line = usable hardware memory

  • Text encoders (FP16)10.59 GiB
  • ComfyUI inference reserve0.80 GiB
  • Safety margin1.37 GiB

Per-stage peak

ComfyUI moves idle models to system RAM between stages; the largest stage sets the requirement. Totals include the safety margin.

  • Text encodePeak12.75 GiB

    Text encoders (FP16) 10.59 + ComfyUI inference reserve 0.80

    Subtotal 11.39 + safety margin (12%) 1.37 ≈ 12.75 GiB

  • Sampling5.84 GiB

    Diffusion model (FP16) 2.64 + Sampling activations 1.77 + ComfyUI inference reserve 0.80

    Subtotal 5.21 + safety margin (12%) 0.63 ≈ 5.84 GiB

  • VAE decode6.99 GiB

    VAE 0.24 + VAE decode activations 5.21 + ComfyUI inference reserve 0.80

    Subtotal 6.24 + safety margin (12%) 0.75 ≈ 6.99 GiB

Generation performance

Denoising steps
50
Conditioning branches per step
1
Latent patch tokens per sample
32,760
Estimated backbone work per step
290.798 TFLOP
Step time at single-device BF16 peak
1,760.28 ms

DiT work estimates cover linear layers and visual self-attention. Enter measured step time and fixed overhead under Sampling and timing.

Compare GPU capacity

Frequently asked questions

How much VRAM does Wan 2.1 T2V 1.3B need?

Our default estimate at the source default resolution, batch 1, FP16 text encoders on GPU and FP16 backbone is 12.8 GiB including the safety margin. These are estimates, not measured guarantees.

Can I offload text encoders?

Yes. The CPU option removes text encoder weights from the GPU estimate, trading GPU memory for system RAM and transfer time. Each component still needs to fit on its assigned device.

Does a green result guarantee compatibility?

No. Green means the estimated memory including safety headroom fits. Backend support, precision support, kernels and runtime allocation can differ. Confirm compatibility and measure your actual workflow.

Where do these estimates come from?

Architecture and file sizes are traced to model sources. The precision table identifies measured GGUF rates, block storage and nominal bit widths. The source data for Wan 2.1 T2V 1.3B was last verified on 2026-11-18. Memory estimates use GiB.