SDXL makes most sense when a project already has illustration checkpoints, LoRAs and conditioning tools built for its architecture. That established ecosystem is the reason to keep it in contention. Compare the correction work it needs with newer models before committing to a new pipeline. Its A tier reflects workflow maturity; image quality still needs a project-specific comparison. Original SDXL 1.0 uses CreativeML Open RAIL++-M, which differs from Stability’s newer Community License. Stability’s current policy prohibits sexual acts and claims broad coverage; its application to older RAIL releases needs separate review. Community checkpoints and hosted generators can add their own licenses and policies.
Commercial use & license
Original SDXL 1.0 uses CreativeML Open RAIL++-M. Do not apply SD3.5 revenue terms or assume a derivative’s added permissions override the base license.
Adult content rules
Original RAIL terms differ from Community licensing. Review the current policy’s sexual-content restrictions and any host’s rules separately.
What you need
Use an SDXL runner with matching encoders and sufficient storage. Measure GPU memory in your selected workflow.
Ranks are recommendations for the task described, based on reviewed documentation. Model ranks cover non-explicit development work. We have not run practical benchmarks. Check content policies and commercial terms separately. How the ranks work →
Sources & further reading +
Last checked 5 October 2026.