How much does Stable Diffusion 3.5 Large Turbo cost via Railwail?

Per-call: €0.04. No monthly minimum, no subscription. Start with €5 free credits.

What is the context window of Stable Diffusion 3.5 Large Turbo?

Stable Diffusion 3.5 Large Turbo supports a unknown context window — enough for typical AI workloads.

How fast is Stable Diffusion 3.5 Large Turbo?

Latency depends on prompt length and load — typically 200ms to 2s for short prompts. We measure p50/p95 in real-time on /rankings.

Is Stable Diffusion 3.5 Large Turbo better than Flux 1.1 Pro Ultra?

It depends on your use case. Stable Diffusion 3.5 Large Turbo (Custom) and Flux 1.1 Pro Ultra (Black Forest Labs) are both strong choices in image generation. Compare them side-by-side at /compare/sd-3-5-large-turbo-stability-vs-flux-pro-ultra.

Stable Diffusion 3.5 Large Turbo

Name: Stable Diffusion 3.5 Large Turbo
Brand: Custom
SKU: sd-3-5-large-turbo-stability
Price: 0.04 EUR
Availability: InStock

Custom

Image Generation

Distilled 4-step variant of SD3.5 Large. 8B params, ~4x faster inference at competitive quality.

Generate with Stable Diffusion 3.5 Large Turbo

Describe what you want and pick a size — the image renders inline.

Size

Result appears here

TL;DR·Last updated May 16, 2026

Stable Diffusion 3.5 Large Turbo is image generation AI model from Custom, priced at €0.000 per 1M input tokens with a unknown context window.

Try Stable Diffusion 3.5 Large Turbo

Prompt

Aspect Ratio

Quality

Direct API access coming soon

Pricing

Price per Generation

Per generation€0.04

API Integration

Use our OpenAI-compatible API to integrate Stable Diffusion 3.5 Large Turbo into your application.

Install

npm install railwail

JavaScript / TypeScript

import railwail from "railwail";

const rw = railwail("YOUR_API_KEY");

const images = await rw.run("sd-3-5-large-turbo-stability", "A beautiful sunset over Tokyo");
console.log(images[0].url);

// Or use the image() method for full control
const res = await rw.image("sd-3-5-large-turbo-stability", "A cat in space", {
  size: "1024x1024",
  n: 1,
});
console.log(res.data[0].url);

Specifications

Price

€0.04

Developer

Custom

Deep dive — Stability AI's Stable Diffusion 3.5 Large Turbo

About Stability AI

Founded 2019 · London, UK

Stability AI was founded in 2019 by Emad Mostaque (CEO until March 2024; succeeded by Prem Akkaraju) and is headquartered in London. The company sponsored and released the original Stable Diffusion (2022, in partnership with CompVis/LMU Munich and Runway), SDXL (2023), SD3 (2024) and SD 3.5 (Oct 2024). After the original Latent-Diffusion team left in mid-2024 to found Black Forest Labs, Stability AI was restructured under new ownership led by Sean Parker. The October 2024 SD 3.5 release included three variants — Large (8.1B), Large Turbo (distilled few-step) and Medium (2.5B) — all under a permissive Community License.

Visit Stability AI →

Architecture

MMDiT distilled to 4 steps via Adversarial Diffusion Distillation

Stable Diffusion 3.5 Large Turbo is the few-step distilled variant of SD 3.5 Large, released October 2024 by Stability AI. The architecture is identical to SD 3.5 Large — an 8.1B parameter Multimodal Diffusion Transformer (MMDiT) with CLIP-L, CLIP-G and T5-XXL text encoders, trained with rectified-flow on a large curated image-text corpus — but Turbo applies Adversarial Diffusion Distillation (Sauer et al. 2023, ADD) to compress the inference schedule from 28-50 steps down to just 4 steps. ADD trains a student model against a teacher U-Net via a hybrid score-distillation + adversarial GAN loss, preserving high-resolution quality while drastically reducing sampling cost. Stability AI reports SD 3.5 Large Turbo achieves quality close to the full Large model at roughly 4-7x faster inference, making it a strong fit for interactive applications and batch generation. The model is open-weights under the same Stability Community License.

Parameters: 8.1B parameters
Context: 256 tokens

What it can do

4-step inference for 4-7x faster generation than SD 3.5 Large
Quality close to full SD 3.5 Large (8.1B MMDiT)
Open weights with permissive Community License
Three text encoders (CLIP-L, CLIP-G, T5-XXL)
1024x1024 native resolution
Compatible with the ControlNet, LoRA and IP-Adapter ecosystem
Self-hostable on consumer GPUs with quantisation
Best for: interactive creative apps, real-time previews, batch pipelines, on-prem chat-with-image.

Training & License

Same pretraining as SD 3.5 Large, followed by ADD distillation with an adversarial loss against the SD 3.5 Large teacher.

License: Stability AI Community License — free for non-commercial and commercial use up to $1M annual revenue; enterprise license required above that threshold.

Known limitations