SageAttention Won't Install in ComfyUI? Fix (2026)

Why Does SageAttention Fail to Install in ComfyUI?

Quick answer: SageAttention installs fail in ComfyUI almost always for one of four reasons: a version mismatch between Python, CUDA, PyTorch, and Triton; a missing Triton include/libs copy; the wrong wheel file for your setup; or an outdated PyTorch build. Fix them in this exact order — Visual C++ → Triton → PyTorch → the SageAttention wheel — and the install succeeds.

SageAttention is an attention kernel — a piece of code that speeds up one of the slowest calculations ComfyUI runs during image and video generation. It does not change what your images look like; it only makes that one calculation faster, which is why the install itself is where almost all the problems happen.

Before jumping into the steps, it helps to know exactly what breaks the install. In order of how often each one shows up:

A version mismatch — Python, CUDA, PyTorch, and Triton all need matching versions. The SageAttention wheel file (a pre-built Python package ending in .whl) is compiled for one specific combination of these — not "any recent version."
Triton's include/libs files not copied into python_embeded — ComfyUI's embedded Python is missing these development files by default, and Triton cannot compile without them.
The wrong wheel downloaded — the ABI tag (the abi3 part of the filename) and CUDA version in the filename must match what you have installed, or pip refuses the install outright.
An outdated PyTorch build — SageAttention needs a PyTorch build compiled against the correct CUDA version, not just any PyTorch you already had installed.

This guide walks through all four in the order that actually works, with the exact commands and version checks at each step.

What You Need Before Installing SageAttention

Supported GPUs

NVIDIA RTX 20xx, 30xx, 40xx, or 50xx series. RTX 50xx requires Triton ≥ 3.3, PyTorch ≥ 2.7, and CUDA ≥ 12.8.

Operating System

Windows 10 or Windows 11 with the ComfyUI Windows portable version already installed.

Required Packages

Visual C++ Redistributable, triton-windows, torch + torchvision + torchaudio, and the SageAttention .whl file.

Version Matching

All four components must have compatible versions — Python, CUDA, Triton, and PyTorch must all match the SageAttention wheel you download.

It also helps to update ComfyUI to the latest version before starting. Older ComfyUI portable builds sometimes ship an outdated bundled PyTorch that conflicts with the CUDA 12.8 build SageAttention needs.

How to Install SageAttention in ComfyUI: Step-by-Step

Step 1 — Install Visual C++ Redistributable

Visual C++ Redistributable provides the runtime libraries that Triton and SageAttention depend on. It must be installed before any other step.

Download the Visual C++ installer.

Download directly from Microsoft: https://aka.ms/vs/17/release/vc_redist.x64.exe

Run the installer.

Double-click the downloaded file and follow the installation steps. Click Install or Repair when prompted.

Restart Windows (fresh install only).

If this is the first time installing Visual C++ on this machine, restart Windows before continuing. If the installer shows "Repair", you already have it installed and no restart is needed.

Visual C++ Redistributable installer — showing the install screen for the Microsoft Visual C++ runtime required for SageAttention — Run the Visual C++ Redistributable installer. If it shows Repair, you already have it.

Step 2 — Install Triton (triton-windows)

Triton is a GPU kernel compiler that SageAttention requires to run. The standard Triton package does not support Windows, so you need to install triton-windows — a Windows-compatible build.

Open the ComfyUI Folder Terminal

Navigate to your ComfyUI_windows_portable folder.

This is the folder that contains the run_nvidia_gpu file — your ComfyUI portable installation root.

Open a terminal in that folder.

Click the address bar in Windows Explorer, type "cmd", and press Enter. A command prompt will open with that folder as the working directory.

Windows Explorer address bar — typing cmd to open a terminal in the ComfyUI_windows_portable folder — Type cmd in the address bar of the ComfyUI_windows_portable folder to open a terminal.

Remove Any Old Triton Installation

Run this command first to remove any previously installed version of Triton:

.\python_embeded\python.exe -m pip uninstall triton

Terminal running pip uninstall triton command in ComfyUI portable folder to remove old Triton installation — Uninstall any previously installed Triton package before installing the Windows version.

Install the Triton Windows Package

Install the Windows-compatible Triton build:

.\python_embeded\python.exe -m pip install -U "triton-windows<3.7"

Terminal running pip install triton-windows command to install the Windows-compatible Triton package for SageAttention — Install triton-windows — the Windows-compatible Triton build required by SageAttention.

Copy Include and Libs Files

This step is required for Triton to function correctly on Windows. You need to download additional include and libs files that match your embedded Python version and copy them into the python_embeded folder.

Check your embedded Python version.

Run the following command to find out which Python version is bundled with your ComfyUI installation:

.\python_embeded\python.exe --version

Terminal showing the embedded Python version in ComfyUI portable — used to select the correct Triton include/libs zip file — Note the exact Python version number — you need it to download the right include/libs zip.

Go to the Triton Windows releases page.

Navigate to the Triton releases on GitHub: https://github.com/triton-lang/triton-windows#8-special-notes-for-embeded-python and find the release version that matches your installed triton-windows version.

Download the matching Python include/libs zip.

Download the zip file that matches your Python version — for example python_3.13.2_include_libs.zip for Python 3.13.11. Make sure the zip matches your exact Python version from Step 1.

Triton Windows GitHub releases page — showing Python include/libs zip files to download for the matching Python version — Download the include/libs zip that matches your embedded Python version from the Triton releases.

Extract and copy the include and libs folders.

Unzip the downloaded file, then copy both the include folder and the libs folder into your python_embeded directory inside ComfyUI_windows_portable.

ComfyUI python_embeded folder — showing the include and libs folders copied from the Triton include/libs zip — Copy the include and libs folders into python_embeded. This step is required for Triton to load.

⚠️ Warning: Skipping the include and libs copy is the single most common reason Triton fails to load. If you see Triton import errors after installation, come back and verify both folders are inside python_embeded.

Step 3 — Install or Update PyTorch

SageAttention requires a specific PyTorch build that matches both your CUDA version and the wheel file you will install. Use the commands below to remove the existing torch and install a fresh build with CUDA 12.8 support.

Remove Old Torch Packages

In the same ComfyUI terminal, run:

.\python_embeded\python.exe -m pip uninstall torch torchvision torchaudio

Terminal running pip uninstall torch torchvision torchaudio command in ComfyUI folder before installing the updated CUDA-compatible build — Uninstall the existing torch packages before installing the CUDA 12.8 build.

Install the CUDA 12.8 PyTorch Build

Install the nightly PyTorch build with CUDA 12.8 support:

.\python_embeded\python.exe -m pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128

Terminal running pip install torch torchvision torchaudio with CUDA 12.8 nightly build for SageAttention compatibility — Install the PyTorch nightly build with CUDA 12.8 support.

💡 Tip: If you have an RTX 50xx GPU, this CUDA 12.8 build is required. For older GPUs like RTX 30xx or 40xx, this build still works well and is the recommended version for SageAttention compatibility.

Step 4 — Download and Install the Correct SageAttention Wheel

Check Your CUDA and Torch Versions

Before downloading the wheel, verify your installed CUDA and PyTorch versions. You need these version numbers to select the correct SageAttention wheel file.

Check CUDA version:

.\python_embeded\python.exe -c "import torch; print(torch.version.cuda)"

Terminal showing the CUDA version reported by PyTorch in ComfyUI embedded Python — used to select the correct SageAttention wheel — Note the CUDA version number — you need it to select the correct wheel file.

Check PyTorch version:

.\python_embeded\python.exe -m pip show torch

Terminal showing pip show torch output with PyTorch version number in ComfyUI — used to match the SageAttention wheel file — Note the PyTorch version from the pip show output.

If the above command does not show version info, use this alternative:

.\python_embeded\python.exe -m pip freeze

Download the Correct SageAttention Wheel

Go to the SageAttention releases page on GitHub: https://github.com/woct0rdho/SageAttention/releases

Find the wheel file that matches your CUDA version, PyTorch version, and Python ABI. These exact filenames are what you'll see on the releases page — match yours to one of these:

sageattention-2.2.0+cu128torch2.10.0andhigher.post4-cp39-abi3-win_amd64.whl

Use this if you have PyTorch ≥ 2.12.0 and CUDA 12.8 (cu128)

sageattention-2.2.0+cu130torch2.9.1andhigher.post4-cp39-abi3-win_amd64.whl

Use this if you have PyTorch <= 2.9.1 and CUDA 13.0 (cu130)

SageAttention GitHub releases page — showing available wheel files to download for different CUDA and PyTorch version combinations — Download the wheel that matches your CUDA and PyTorch versions from the SageAttention releases page.

Install the Wheel File

Move the wheel file into your ComfyUI_windows_portable folder.

Copy or move the downloaded .whl file into the same folder where you have been running all the commands (the root ComfyUI_windows_portable folder).

Install the wheel using pip.

Run the install command with your exact wheel filename. Replace the filename below with the one you downloaded:

.\python_embeded\python.exe -m pip install ./sageattention-<your-file-name>.whl

Alternatively, type .\python_embeded\python.exe -m pip install ./sageattention and press Tab to auto-complete the filename.

Terminal running pip install sageattention wheel file command in ComfyUI_windows_portable folder — Install the SageAttention wheel using pip. The filename must match exactly.

⚠️ Warning: If pip fails with an error, the most common cause is a version mismatch — your CUDA, PyTorch, or Python version does not match what the wheel requires. Double-check your versions from Steps 3 and 4 and download the wheel that matches.

SageAttention's speed boost is most noticeable in heavier workflows such as FLUX workflows — the larger the model, the more time the attention calculation takes, and the more time SageAttention saves.

Step 5 — Verify SageAttention Is Working

After installation, start ComfyUI and verify that SageAttention and Triton are loading correctly.

Start the ComfyUI server.

Launch ComfyUI normally using the run_nvidia_gpu file.

Watch the console output.

Look for messages about SageAttention or Triton loading at startup. Successful installation will show confirmation messages in the console.

Load a workflow and run it.

Load a simple workflow that uses attention operations or the SageAttention node. If the workflow runs without errors, installation was successful.

💡 Tip: If you see import errors or Triton errors in the console, check that the include and libs folders from Step 2 are correctly placed inside python_embeded, and that all version numbers match.

If ComfyUI flags any missing nodes when you load the test workflow below, follow the installing missing custom nodes guide to resolve them before continuing.

Working Setup Example

For reference, here is a confirmed working configuration:

OS:Windows 11

GPU:RTX 4090

Python:Python 3.13.11 (embedded)

Triton:triton-windows 3.5.1.post24

PyTorch:torch 2.12.0.dev20260219+cu128

SageAttention wheel:sageattention-2.2.0+cu128torch2.9.0andhigher.post4-cp39-abi3-win_amd64.whl

Does SageAttention Slow Down or Change Image Quality?

No. SageAttention only speeds up the attention calculation — it does not change the output. Once it's installed correctly using the steps above, you should see a real drop in generation time with no visible difference in the result.

In most cases, generation time drops by 30–50%, and sometimes more depending on the model and hardware. For example, an image that normally takes 115 seconds without SageAttention can take only 45 seconds with it enabled — close to 70% faster.

ComfyUI generation with SageAttention off — 115 seconds — SageAttention OFF — 115 seconds

ComfyUI generation with SageAttention on — 40 seconds — SageAttention ON — 40 seconds

💡 Tip: FLUX-based models typically see the biggest speed gains. Try a longer generation — more steps or a larger resolution — to see the difference more clearly.

Why Isn't SageAttention Working After Installation?

⚠ Triton fails to import or load

The include and libs folders were not copied into python_embeded. Go back to Step 2 and copy both folders from the Triton releases zip that matches your Python version.

⚠ pip fails when installing the SageAttention wheel

Version mismatch between the wheel and your installed CUDA / PyTorch / Python versions. Re-run the version checks from Step 4 and download the wheel that exactly matches your setup.

⚠ SageAttention installs but does not appear to speed anything up

Verify that the wheel ABI tag (cp39-abi3) is compatible with your embedded Python version. Also confirm that SageAttention is actually being called in your workflow nodes.

⚠ Black or corrupted frames on RTX 50xx GPUs

This is a known issue on Blackwell (RTX 50xx) cards: SageAttention's fp8 kernel can overflow fp16 on models that run fp8 weights with fp16 math, producing NaN values that show up as black frames. It shows up most often at 2:1 resolutions like 1280×640. If you hit this, try a standard resolution first (1024×1024 or 1280×720) before assuming the install itself is broken.

⚠ All steps fail — nothing works

Reinstall the ComfyUI Windows portable version from scratch and follow all steps again from the beginning. Make sure your GPU is supported (RTX 20xx or newer).

If the problem you're seeing isn't listed above, it may not be specific to SageAttention. Check the general ComfyUI troubleshooting guide for node errors, dependency conflicts, and other setup issues unrelated to attention kernels.

Frequently Asked Questions

SageAttention in ComfyUI supports NVIDIA RTX 20xx, 30xx, 40xx, and 50xx series GPUs. If you have an RTX 50xx GPU, you need Triton 3.3 or higher, PyTorch 2.7 or higher, and CUDA 12.8 or higher.

To install SageAttention in ComfyUI on Windows you need: Visual C++ Redistributable, the Triton package for Windows (triton-windows), PyTorch (torch, torchvision, torchaudio) with the matching CUDA version, and the ComfyUI Windows portable version already installed.

Download the wheel from the SageAttention GitHub releases page. Match the wheel to your CUDA version (e.g. cu128 for CUDA 12.8) and your PyTorch version. For example, if you have PyTorch 2.9 or higher with CUDA 12.8, choose the wheel named sageattention-X.X.X+cu128torch2.9.0andhigher-cp39-abi3-win_amd64.whl.

The most common causes are: a version mismatch between Python, PyTorch, Triton, and the wheel file; forgetting to copy the Triton include and libs folders into python_embeded; using the wrong torch build for your CUDA version; or selecting a wheel with an ABI tag incompatible with your embedded Python version.

Yes. After installing the triton-windows package, you must download the Python include/libs zip from the Triton Windows releases page that matches your embedded Python version, then copy the include and libs folders into your python_embeded folder. Skipping this step is one of the most common reasons Triton fails to load.

No. SageAttention speeds up the attention calculation during generation without changing the output. Images and videos generated with SageAttention on or off are visually identical — only the generation time changes.

What to Do Next

Once SageAttention is confirmed working, put it to use in a real workflow — the ComfyUI roadmap lays out what to learn next, starting with the FLUX workflows where the speed gain matters most.