SAVRN Model Hub · Comparisons
sdxl-turbo vs stable-diffusion-xl-base-1.0
Sdxl-turbo has 2.6B parameters and stable-diffusion-xl-base-1.0 has 2.6B parameters; sdxl-turbo is released under other and stable-diffusion-xl-base-1.0 under Open RAIL++ License; at 16-bit, sdxl-turbo needs about 6.2 GB (1x MI300X from $1.85 an hour) and stable-diffusion-xl-base-1.0 about 6.2 GB (1x MI300X from $1.85 an hour).
| Field | sdxl-turbo stabilityai/sdxl-turbo | stable-diffusion-xl-base-1.0 stabilityai/stable-diffusion-xl-base-1.0 |
|---|---|---|
| Publisher | Stability AI | Stability AI |
| Task | Text to image | Text to image |
| Modality | Image | Image |
| Parameters, as reported | 2.6B parameters | 2.6B parameters |
| Architecture | Not stated | Not stated |
| Library | diffusers | diffusers |
| Context length | Not stated | Not stated |
| Repository size | 55.5 GB | 76.9 GB |
| Artifact formats | safetensors, onnx | safetensors, onnx |
| License | other | openrail++ |
| Access | Open weights, no gate | Open weights, no gate |
| Memory at 16-bit (weights and margin) | 6.2 GB | 6.2 GB |
| Cheapest GPUs at 16-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Memory at 4-bit (weights and margin) | 1.5 GB | 1.5 GB |
| Cheapest GPUs at 4-bit, per hour | 1x MI300X, $1.85 | 1x MI300X, $1.85 |
| Revision viewed | 71153311d3db | 462165984030 |
| Downloads reported by the hub | 790.4k | 2.9M |
| Last observed | 2026-09-18 | 2026-09-18 |
An evaluation row appears only where at least two of these models report the same benchmark with the same stated configuration, metric, unit and setup. Different evaluators stay named in each cell. Values are shown as reported: no unit conversion, no ranking.
SAVRN's Notes on sdxl-turbo
Stability AI distilled this 2.6-billion-parameter text-to-image model from SDXL 1.0 so it samples in 1 to 4 steps, and with so few steps per image, throughput per card rather than memory sizes the deployment. At 16-bit it needs 6.2 GB, at 8-bit 3.1 GB, at 4-bit 1.5 GB, and the cheapest configuration we track, a single MI300X with 192 GB at $1.85 an hour, sits mostly idle unless you batch requests onto it. The repository itself is 39 files totaling 55.5 GB.
The license is listed as other, not a standard open-source grant, and the publisher points commercial users to stability.ai/license. Read that page before anything goes into production; the terms of a commercial deployment are set there, not in the model files. Released November 27, 2023 and last updated July 10, 2024, with no configuration record in our data, your own prompts on your own hardware are the check.
SAVRN's Notes on stable-diffusion-xl-base-1.0
The repository is 76.9 GB across 57 files; the run is 6.2 GB. One 16-bit copy of this 2.6B-parameter text-to-image model is 5.1 GB of weights, 4-bit needs 1.5 GB, and the page's 63.8 GB of weight files in safetensors and onnx hold far more than either, so pull only the format you serve. One MI300X at $1.85 an hour on-demand is the cheapest listing on our Index that fits, and this fills one thirtieth of its 192 GB.
Open RAIL++ permits commercial use but carries use-based restrictions that pass to your downstream users, so they belong in your own terms if you sell images; nothing is gated. Stability AI pairs this base with a separate refiner model for the final denoising steps and says the base runs standalone; decide which you are deploying, since the refiner is a second download and footprint. The files last moved 2023-10-30.
Questions
Which is larger, sdxl-turbo or stable-diffusion-xl-base-1.0?
sdxl-turbo (2.6B parameters) is larger than stable-diffusion-xl-base-1.0 (2.6B parameters), by the parameter counts their publishers report.
Which is cheaper to run, sdxl-turbo or stable-diffusion-xl-base-1.0?
At 4-bit, sdxl-turbo fits on 1x MI300X from $1.85 an hour and stable-diffusion-xl-base-1.0 on 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use stable-diffusion-xl-base-1.0 commercially?
Yes, with conditions. stable-diffusion-xl-base-1.0 is released under Open RAIL++ License. Open RAIL++ permits use, including commercial use, subject to the use-based restrictions listed in the license, which must be passed on to downstream users.