Leonardo AI
T1 — CLIP-BasedFast iteration and asset creation for creatives.
How Leonardo AI reads prompts
Leonardo AI is classified as CLIP-Based (T1) — weighted keywords with clip tokenisation. CLIP encoder with double-colon weight syntax term::weight. "Elements" = LoRAs (max 4, combined weight ≤1.00). ~1,000-char limit. Front-loading emphasis critical.
Prompt tips
Use weighting for emphasis (concept::1.5). Choose appropriate model for style.
Why prompt optimisation matters
Critical for game/concept art quality and style consistency. Promagen's Prompt Lab automatically formats your selections into Leonardo AI's native prompt structure.
Negative prompt support
Leonardo AI has a separate negative prompt field — you can specify elements to exclude in a dedicated input. This is the most powerful form of negative prompt support.
Full negative prompt support guide →Platform notes
SD-based with multiple fine-tuned models; CLIP 77-token limit applies.
Example prompt
Frequently asked questions
What is the character limit for Leonardo AI?
Leonardo AI accepts prompts up to 1000 characters. The ideal writing range is 250–400 characters (around 100 characters is the sweet spot where the platform produces its best results).
Does Leonardo AI support negative prompts?
Yes. Leonardo AI has a separate negative prompt field where you can specify elements to exclude from the generated image.
How should I write prompts for Leonardo AI?
Leonardo AI uses CLIP-Based prompt format (T1). Weighted keywords with CLIP tokenisation. Use weighting for emphasis (concept::1.5). Choose appropriate model for style.
Other CLIP-Based platforms
These platforms share the same prompt architecture — prompts written for one will generally work well on the others.