Leonardo AI Model Selection Reference (2026 Interface)

Leonardo AI Model Selection Reference (2026 Interface)

Leonardo's documented web flow offers two distinct choices: enter a prompt on the home page and let Leonardo select a model, or open Image Creation and select a model explicitly. This reference helps make that choice without treating a model name as a promise about an image.

Interface documentation dated: 2026-02-17. The official image-creation help article carries that modification date. Catalog snapshot checked: 2026-07-16. The catalog names below were present in Leonardo's official Commonly Used Models page on that date.

Documented: the home-page automatic route, the explicit Models menu in Image Creation, the Fast and Ultra generation-mode labels, the model-dependent settings caveat, and the provider/model names in the catalog.

Locally checked: source URLs, catalog spellings, table entries, and the text-only worksheet were checked for internal consistency on 2026-07-16.

Not tested: no signed-in interface session, image generation, model comparison, or output assessment was performed. No independent image-generation comparison was performed for this guide.

Choose automatic or explicit selection first

The official image-creation article describes both paths:

1. Automatic selection: enter a prompt in the home-page prompt box and choose Generate. Leonardo says it selects a model based on that prompt.
2. Explicit selection: expand the prompt box to expose more controls or enter Image Creation from the Image option. In Image Creation, use the Models menu to choose an image model.

Use automatic selection when the model identity is not part of the record you need to keep. Use explicit selection when a workflow, handoff, or troubleshooting note must name the model. That is a reproducibility distinction, not a quality ranking.

The Image Creation sidebar also documents Fast and Ultra as generation modes. This guide records only that those two labels are documented. It does not make price, token, speed, fidelity, or output comparisons between them.

Task-to-model catalog worksheet

The catalog calls itself a snapshot of commonly used models and warns that it is not comprehensive. Leonardo's official AI Image Models article supplies the application wording below. The table reframes those applications as intents to investigate, not as rankings or outcome claims.

The Commonly Used Models catalog confirms Flux Dev under Black Forest Labs and Nano Banana Pro under Google. It also confirms all three Leonardo.Ai entries above. Other current or legacy models may exist; absence from this shortlist is not an availability claim. The application wording identifies a candidate to inspect, while the worksheet below controls the actual comparison.

Settings vary by model

Leonardo's Image Creation article explicitly warns that available settings vary from model to model. Treat a model change as a settings-schema change:

• re-read the visible controls after changing the model;
• do not copy a hidden or unavailable setting from a previous model into the new run notes;
• record the chosen model before recording dimensions, generation mode, style, or reference-image controls;
• if a control disappears, describe it as unavailable in that observed interface state rather than unsupported everywhere.

This prevents a worksheet from implying that every catalog model accepts the same controls.

Prompt-control worksheet

Complete one row before changing the model. The final column identifies the single field allowed to change in the next comparison.

The table is reproducible without generating an image: compare the two rows character by character. Prompt, dimensions, mode, and style/reference state remain fixed; only the exact model label changes.

For a live run, replace “record visible value” with what the interface actually shows. If switching the model changes which settings are available, stop and annotate that difference before continuing. Do not silently fill a missing control with the previous model's value.

One-variable troubleshooting

1. Copy the current row and give it a new run label.
2. Keep the prompt text unchanged.
3. Change only automatic/explicit selection or the exact model label.
4. Re-check the sidebar because controls may vary by model.
5. Record any settings-schema difference before changing dimensions, mode, or style.
6. If the issue concerns an unwanted element, hold the model steady and move to the Leonardo negative-prompt troubleshooting reference.

This sequence isolates selection from prompt wording. It cannot establish which model is preferable; that would require owned comparison evidence outside this guide's boundary.

Primary sources

• How to Generate Images with Leonardo.Ai — official help article, modified 2026-02-17; automatic and explicit selection, Image Creation controls, generation-mode labels, and the settings-vary-by-model warning.
• Commonly Used Models — official provider/model catalog snapshot checked 2026-07-16.
• AI Image Models — official Leonardo model-family overview supplying the documented application examples used in the task table; its ranking-style labels are not repeated here.