Nvidia Sana: solid text-to-image with good composition and detail—reliable baseline for a wide range of prompts.
The best image results come from specific composition, style, and lighting language. Be explicit about what should be in frame and what should feel dominant.
Best results start with a precise subject, composition, and style direction.
Sana Base on Pixio is Nvidia Sana's base text-to-image model: solid quality, good composition and detail, and a reliable baseline for a wide range of prompts. Use it when you want Sana output without the extra quality of v1.5 or the speed focus of Sprint.
Sana Base on Pixio is Nvidia Sana's base text-to-image model: solid quality, good composition and detail, and a reliable baseline for a wide range of prompts. Use it when you want Sana output without the extra quality of v1.5 or the speed focus of Sprint.
| Mode | Input | Best for |
|---|---|---|
| Text to Image | Prompt only | Scenes, characters, products from a single prompt |
| Option | Values | Notes |
|---|---|---|
| Aspect ratio | 1:1, 16:9, 9:16 (check Pixio) | Match deliverable |
| Credits | Plan-based | Check model card in Pixio |
Credits are plan-based; check the model card in Pixio for your plan and cost per image.
[Subject] + [Composition] + [Lighting] + [Style]. One clear concept per prompt; be specific about pose, setting, and mood.
"Portrait of a woman in a park. Natural daylight. Photoreal, relaxed."
"A coffee cup on a wooden table. Morning light. Simple, cozy."
"City street at night. Neon signs. Urban, moody."
"Fantasy knight in armor. Dramatic lighting. Illustration style."
| Scenario | Best choice |
|---|---|
| Sana baseline, general use | Sana Base |
| Sana best quality | Sana v1.5 |
| Sana fastest | Sana Sprint |
| Sana fast (v1.5) | Sana v1.5 Fast |
| Non-Sana | Flux Pro, Imagen 4 |
Tell the model what should dominate the frame first.
Use lighting language early; it changes everything downstream.
When editing, describe what stays, not just what changes.
References help when continuity matters more than novelty.
A strong image prompt defines the subject, composition, lighting, and finish instead of leaving them implied.
Use precise visual language to control subject, composition, lighting, and style from the start.
Preserve the useful parts of the image while steering the rest with masks, references, or prompt edits.
Bring in reference images or LoRAs when consistency is more important than exploration.
Sana Base is strongest when the visual brief is specific about framing, style, and what should read first.
Use it for campaign images, product shots, subject consistency, or polished concept work.
When editing, say exactly what changes and what must remain untouched.
Lock the subject, composition, and lighting direction before you chase style nuance.
Use references or edits when the same subject, style, or layout has to survive across versions.
Once the frame works, refine only the weak areas instead of rewriting the whole composition.
Finish strong compositions by scaling them without rebuilding the frame from scratch.
Use editing tools after the initial generation when the composition is right but the details still need polish.
