Baidu · Text to image
ERNIE Image Turbo
ERNIE-Image Turbo is Baidu’s distilled, low-step image variant for quicker visual iteration. It prioritizes throughput and appealing aesthetics over the stronger general instruction fidelity of the standard ERNIE model. Use it to compare mood-board directions, social-image concepts, editorial hero images and product-scene drafts, then move to a more controlled model when a final layout has many exact requirements.
Generation mode
Text to image
Credit cost
1–2 credits
Observed generation time
Not enough public data
Release date
Apr 15, 2026
Current credit pricing
| Duration | Configuration | Credits |
|---|---|---|
| 0 seconds | 1024_8 | 1 |
| 0 seconds | 2048_8 | 2 |
| 0 seconds | 512_8 | 1 |
Availability and credit rows come from Flowctory's active production catalog. Observed time is shown only after at least five completed generations and is not a controlled benchmark. Model availability, limits, and costs can change.
Use model