> ## Documentation Index > Fetch the complete documentation index at: https://comfyui-mcp.artokun.io/docs/llms.txt > Use this file to discover all available pages before exploring further. # Z-Image in ComfyUI: The Low-VRAM Speed King (2026) > Run Z-Image Turbo and Base — Alibaba's 6B text-to-image model — locally in ComfyUI under 8 GB VRAM. Install, GGUF tips, LoRA training, and the V2 ULTRA workflow. *by [artokun](https://github.com/artokun) · June 16, 2026 · z-image · image · ComfyUI · model highlight* Most "run it locally" image models come with a catch: a 24 GB card, a 50 GB download, or a five-second wait per frame. **Z-Image** quietly breaks all three. It's a **6B-parameter** text-to-image model from **Alibaba's Tongyi-MAI lab** that, as a GGUF quant, generates photoreal 1024px images **under 8 GB of VRAM** — and on a 4090 it does it in roughly **2.3 seconds at 8 steps**. This is the low-VRAM speed entry in our model-highlight series, and it earns the title. Below: what Z-Image actually is, the **Turbo vs Base** split (fast vs trainable), the V2 ULTRA workflow extras (ControlNet, Detail Daemon, an 8-step LoRA, and the fix for Z-Image's infamous "same seed, same image" problem), how to **train your own Z-Image LoRAs**, where it beats — and loses to — the alternatives, and the fastest way to get it running: a one-command install with [comfyui-mcp](https://github.com/artokun/comfyui-mcp) and the [Panel](../panel), instead of hand-downloading GGUFs into the right folders. > **TL;DR — one-command setup.** Install comfyui-mcp, apply the > [`z-image-turbo`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-turbo) pack (fast) or > [`z-image-base`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-base) pack (trainable) — nodes + GGUF models land > in the right folders, validated — and drive the graph from your own Claude session > through the [Panel](../panel). Jump to [Install](#install-z-image-in-comfyui). ## What is Z-Image? Z-Image (released by **Alibaba's Tongyi-MAI lab**; Turbo landed November 27, 2025, Base in January 2026) is a **\~6B-parameter** text-to-image model built on a **Scalable Single-Stream DiT (S3-DiT)** architecture — text tokens, visual semantic tokens, and image VAE tokens are concatenated into one unified input stream, which is a big part of why it stays so parameter-efficient. It uses a **Qwen3-4B** text encoder (not CLIP-L/T5), and the **same `ae.safetensors` VAE as Flux**. The headline isn't size — it's the efficiency-to-quality ratio. With only 6B params, Tongyi claims output **comparable to closed-source flagships in the 20B+ range**, and reviewers consistently call out its **photorealistic portraits** and solid **English + Chinese text rendering**. A fair flag: that "rivals 20B models" line is a vendor framing, scoped most strongly to **photoreal people and general scenes** — treat it as "punches well above its weight," not a universal SOTA claim. What makes it the **low-VRAM speed king**: as a GGUF quant the diffusion model fits in **under 8 GB**, and Turbo needs only \~8 steps. People are running it on 6 GB cards. That's the pitch — flagship-ish quality on hardware that can't touch Flux.2 or a 9.3B [Ideogram 4](./ideogram-4-comfyui). ## Z-Image vs the alternatives There's no single "best" — pick by the job: | Pick… | When you need… | | ----------------- | ------------------------------------------------------------------------------------------- | | **Z-Image Turbo** | **Speed + low VRAM** — fast ideation, batch iteration under 8 GB | | **Z-Image Base** | **Trainable** photoreal base, diverse styles, LoRA finetuning | | **Ideogram 4** | Readable **in-image text**, posters, logos, layout control | | **Flux.2** | Top-tier **photorealism** and a mature production default | | **Qwen-Image** | Text + multilingual on a larger dev model — see our [Qwen-Image post](./qwen-image-comfyui) | Where Z-Image wins: **VRAM and speed**, full stop — nothing else here runs flagship- ish quality in 6–8 GB at this pace, and Base is genuinely cheap to finetune. Where it loses: in-image **typography and layout** go to [Ideogram 4](./ideogram-4-comfyui) (a 9.3B model purpose-built for text), and the last 5% of large-scene **photorealism** still favors Flux.2. The text-and-layout deep-dive lives in the [Ideogram 4 post](./ideogram-4-comfyui); Z-Image earns this spotlight for flagship-ish output on hardware the others can't touch. ## System & VRAM requirements (GGUF) Z-Image is distributed here as **GGUF** quants, loaded through [ComfyUI-GGUF](https://github.com/city96/ComfyUI-GGUF) (`UnetLoaderGGUF` / `CLIPLoaderGGUF`). The quant tier is your VRAM lever: | Quant | VRAM (rough) | Notes | | ------------ | -------------- | --------------------------------------------------- | | **Q5\_K\_S** | **under 8 GB** | The low-VRAM pick | | **Q8\_0** | \~12–16 GB | Best quality; what both V2 workflows ship pinned to | | BF16 (full) | \~14–16 GB | Top-tier consumer cards | **Heads-up on quants:** the installer menu offers a **Q6\_K** option, but **no Q6\_K GGUF exists on the Aitrepreneur/FLX mirror** — use **Q5\_K\_S** (under 8 GB) or **Q8\_0** (best quality) instead. The shipped workflows pin Q8\_0; to drop to the low-VRAM tier, just point the `UnetLoaderGGUF` widget at the `Q5_K_S` filename (or uncomment that tier in the pack's `manifest.yaml`). The text encoder is `Qwen3-4B-UD-Q6_K_XL.gguf` and the VAE is the Flux `ae.safetensors`. ## Install Z-Image in ComfyUI The manual route works: install [ComfyUI-GGUF](https://github.com/city96/ComfyUI-GGUF) plus the workflow's other custom nodes, download the GGUF UNet, Qwen3-4B encoder, VAE, ControlNet patch, and upscaler into the right folders, then load the workflow. That's a lot of click-and-place. ### The fast way — comfyui-mcp + the Panel The [comfyui-mcp](https://github.com/artokun/comfyui-mcp) packs remove all of it. One declarative manifest installs every custom node (ComfyUI-GGUF, rgthree, KJNodes, Detail Daemon, controlnet\_aux, and more) and pulls each model to the correct folder — and the same manifest drives both an MCP-native install and the generated double-click scripts: ```bash theme={null} # Z-Image Turbo (fast, under 8 GB) — MCP-native from a Claude Code session apply_manifest --path packs/z-image-turbo/manifest.yaml # Z-Image Base (trainable, more styles) apply_manifest --path packs/z-image-base/manifest.yaml # or one-click, run from your ComfyUI root packs/z-image-turbo/install-windows.bat # Windows packs/z-image-turbo/install-runpod.sh # RunPod / Linux ``` Want to compare trained LoRAs? Add the [`z-image-xy-plot`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-xy-plot) pack. Run the generated installer from your **ComfyUI root** (the folder containing `custom_nodes/` and `models/`), then restart ComfyUI (or use ComfyUI-Manager → *Install missing custom nodes*) and load `workflow.json`. Because the packs ship with the [plugin](../plugin), your **own Claude session can drive the live graph through the [Panel](../panel)** — swap quant tiers, wire the ControlNet, toggle the Detail Daemon, and iterate on prompts conversationally, with full Ctrl+Z undo and no API keys. Every model URL is CI-validated for reachability and size, so a link never quietly rots ([here's why that matters](./installer-packs-that-cant-rot)). ## Turbo vs Base — pick by the job Z-Image ships in two flavors, and the difference is the whole story: | | **Z-Image Turbo** | **Z-Image Base** | | --------------- | --------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | | Built for | **Speed** | **Quality + training** | | Distillation | Decoupled-DMD distilled | Non-distilled foundation model | | Steps | **\~4–8** (sub-second on data-center GPUs) | **\~30–50**, CFG 3–5 | | Negative prompt | Not effective (CFG baked in) | **Supported** at CFG > 1 | | Best at | Fast ideation, batch iteration | Diverse styles, finetuning, LoRA training | | Pick the pack | [`z-image-turbo`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-turbo) | [`z-image-base`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-base) | **Turbo** is the one you reach for first: \~4–8-step sampling, sub-second latency on beefy GPUs, \~2.3s for 1024px on a 4090. It's distilled, so it bakes CFG in — meaning **negative prompts don't really work**; you steer with the positive prompt instead. **Base** is the non-distilled foundation model. It costs you steps (30–50) and time, but it gives back a **higher artistic ceiling, more diverse styles, real negative prompts, and — critically — it's the finetuning-friendly variant**. If you're training LoRAs, you train on Base. The handy part: **a LoRA trained on Base generally applies to the Turbo workflow too** (verify on your own LoRA), so you can train once and deploy on the fast path. ### The V2 ULTRA workflow extras Both packs ship a tuned **V2 ULTRA** workflow that's a lot more than a bare KSampler. Highlights: * **Fun-ControlNet-Union** — one ControlNet patch covering Canny, HED, Depth, Pose, and MLSD (loaded via `ModelPatchLoader` → the Z-Image Fun ControlNet node). Recommended strength \~0.65–0.80. Base uses the `Z-Image-Fun-Controlnet-Union-2.1` weights; Turbo uses `Z-Image-Turbo-Fun-Controlnet-Union` fp8. * **Detail Daemon** ([ComfyUI-Detail-Daemon](https://github.com/Jonseed/ComfyUI-Detail-Daemon)) — a detailer pass that adds fine texture and skin/fabric detail without re-rolling the composition. * **8-step distill LoRA** (Base pack) — `Z-Image-Fun-Lora-Distill-8-Steps` lets the Base model run at Turbo-like step counts when you want speed from the trainable model. * **SeedVarianceEnhancer** (Turbo pack) — the fix for Z-Image's most-complained-about quirk. Turbo's early sampling step **over-determines** the image, so different seeds produce near-identical results ("same seed, same image" — really "every seed, same image"). SeedVarianceEnhancer injects controlled, prompt-related noise into the text embeddings between the encoder and the sampler, restoring real seed-to-seed diversity. (Per the [upstream issue](https://github.com/Tongyi-MAI/Z-Image/issues/16) this is a known trait of the distilled model, not a bug in your graph.) * txt2img / img2img / inpainting branches with **rgthree group toggles**, plus a `4x-ClearRealityV1` upscale path. ### Train your own Z-Image LoRAs Z-Image's low-VRAM advantage carries straight into **training**. Because it's a \~6B **single-stream** model (no WAN-style hi/lo expert split), it's the lightest LoRA-training target in our stack — a 4090 is comfortable, and smaller cards work with float8 quantization at 512–768 resolution. Use the **[`ai-toolkit-trainer`](../plugin) skill** (ostris **AI-Toolkit**, MIT — a standalone web-UI trainer, not a ComfyUI node) to train. Sensible starting points for Z-Image: | Param | Starting point | | ----------------- | ----------------------------------------------- | | Linear rank / dim | 16–32 (32 for detailed characters/styles) | | Learning rate | 1e-4 (5e-5 for tighter identity) | | Steps | 1500–3000 (dataset-dependent) | | Resolution | 768 (or 1024 — Z-Image's native range) | | Multi-stage | **OFF** (single-stream model, not WAN's MoE) | | Optimizer / Quant | AdamW8bit / float8 (enables sub-12 GB training) | *(These are aggregated community starting points, not read from AI-Toolkit's `config/examples/*.yml` — open the actual example config in your clone and tune.)* **Train on Base, deploy anywhere.** Then drop the resulting `.safetensors` into `models/loras/` and compare your candidates side by side with the [`z-image-xy-plot`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-xy-plot) pack — a utility workflow that renders a labelled XY grid of multiple trained LoRAs (and/or checkpoints) at once. Bring your own LoRAs; that pack installs none. ## What to make with it * **Photoreal portraits & fashion** — Z-Image's standout strength; the go-to for faces, headshots, and editorial looks at low VRAM. * **Fast batch ideation** — Turbo's speed makes it ideal for spinning dozens of concepts, then refining the keepers on Base. * **Custom characters & styles** — train a LoRA on Base, deploy on Turbo. * **Guided composition** — Fun-ControlNet for pose/depth/canny-driven scenes. ## Settings that matter The packs ship tuned, but for reference (community-tested presets): | Use | Steps | CFG | Sampler | Scheduler | | -------------------------- | ----- | --- | ----------------- | --------- | | **Turbo — sharpest** | 10 | 1.0 | `dpmpp_sde` | `beta` | | **Turbo — author pick** | 14 | 1.0 | `res_2s` | `simple` | | **Turbo — beauty/fashion** | 10 | 1.0 | `euler_ancestral` | `beta` | | **Base — full quality** | 22–30 | 4.0 | `res_2s` | `beta` | * **Prompt in natural language**, not tags — the Qwen3 encoder is an LLM, so "Professional headshot of a confident businesswoman, soft studio lighting, sharp focus on the eyes, Canon EOS R5" beats "masterpiece, best quality, 1girl." * **Turbo has no working negative prompt** — steer with the positive prompt. * **Native resolutions:** 1024×1024 (or 1328×1328 native), 896×1152, 832×1216, 1280×720. **Dimensions must be divisible by 16.** * **ControlNet** pairs best with `res_2s` / `res_5s` + a `beta57` scheduler. ## Troubleshooting * **"Torch not compiled with CUDA enabled."** A CPU-only torch is installed. Reinstall the matching CUDA torch wheel (e.g. cu126/cu128) into the ComfyUI `python_embeded` / venv. * **Pascal / Maxwell cards (GTX 10xx / 9xx) fail or fall back to CPU.** Newer cu12x wheels dropped those compute capabilities — reinstall the **cu126** torch build from `https://download.pytorch.org/whl/cu126`. * **No Q6\_K file found.** Expected — there's no Q6\_K on the mirror. Use **Q5\_K\_S** or **Q8\_0**. * **Every seed produces the same image (Turbo).** A known trait of the distilled model. Use the **SeedVarianceEnhancer** node (shipped in the Turbo pack) to restore diversity, or generate on Base. * **Missing custom nodes after install.** Restart ComfyUI, or use ComfyUI-Manager → *Install missing custom nodes* to pull node Python deps. * **ControlNet pack: `DWPreprocessor` (DWPose) doesn't appear in the node list.** controlnet\_aux's DWPose silently fails to register when **`scikit-image`** is missing (it's in the node's `requirements.txt`, so a proper install resolves it — the pack also pins it). Install `scikit-image` into the ComfyUI venv and restart. The render-verified ControlNet packs (Turbo + Base) take a pose/canny/depth from your input image via DWPose and guide a fresh generation. * **OOM.** Drop to the **Q5\_K\_S** GGUF, lower the resolution, and confirm you've set `clear_vram` before switching from another model family. ## FAQ **Is Z-Image open source?** The weights are openly published by Tongyi-MAI on HuggingFace and widely usable locally. Confirm the exact license terms on the model card for your use case (*unverified here* — check before commercial use). **Turbo or Base — which should I install?** Turbo for speed and fast iteration under 8 GB; Base if you want maximum quality, diverse styles, working negative prompts, or you're training LoRAs. **How much VRAM do I really need?** The **Q5\_K\_S** GGUF runs **under 8 GB**; Q8\_0 wants \~12–16 GB. People run it on 6 GB cards with low quants. **How fast is it?** Turbo does a 1024px image in roughly **2.3 seconds at 8 steps** on an RTX 4090; sub-second on data-center GPUs. **Can I train a Z-Image LoRA, and on which model?** Yes — train on **Base** with the [`ai-toolkit-trainer`](../plugin) skill; a Base-trained LoRA generally works in the Turbo workflow too. **Why don't negative prompts work in Turbo?** Turbo is DMD-distilled with CFG baked in, so the negative conditioning has little effect. Use Base for real negative prompts, or steer Turbo via the positive prompt. **Why is there no Q6\_K download?** The installer menu lists it, but no Q6\_K GGUF is published on the Aitrepreneur/FLX mirror — use Q5\_K\_S or Q8\_0. *** ## Get it running in one command 1. Install [comfyui-mcp](https://github.com/artokun/comfyui-mcp) and the [Panel](../panel) — the panel auto-starts a background agent on your Claude subscription (no API keys; sign in with `claude` once). 2. Apply the [`z-image-turbo`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-turbo) pack (fast) or [`z-image-base`](https://github.com/artokun/comfyui-mcp/tree/main/packs/z-image-base) pack (trainable) — nodes + GGUF models land in the right folders, validated. 3. Open the [Panel](../panel) and let the panel's agent wire the ControlNet, swap quant tiers, and iterate on prompts for you. That's the whole point of the project: expert ComfyUI setups that install in one step and drive themselves from your own agent session. **Next in the series:** [ERNIE-Image](./ernie-image-comfyui) — the Apache-2.0 text powerhouse with clean multilingual EN/CN/JP type.