Tasks and inputs
ERNIE Image Turbo is listed for Text-to-image. It accepts text and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Oct 7, 2026
ERNIE Image Turbo is Baidu's Apache-2.0 text-to-image checkpoint for locally generated visual assets.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
ERNIE Image Turbo is Baidu's Apache-2.0 text-to-image checkpoint for locally generated visual assets.
Turbo is the distilled release and the publisher recommends eight inference steps. Begin with those settings, then measure whether the shorter generation meets the brief. Increasing steps without checking the recommended guidance can waste compute without improving the result.
Use a brief with exact labels, object counts and spatial relationships. For a poster, check every word and the reading order. For a multi-panel asset, verify that the subject and visual style remain consistent across panels. Attractive lighting cannot compensate for a misspelled product name.
Record whether prompt enhancement is enabled. An enhanced prompt and an unmodified prompt are different inputs, so compare them explicitly rather than attributing every change to the checkpoint alone.
Use the named repository with the documented pipeline and a compatible runtime. Verify the memory required by your chosen dimensions and precision before promising support for a particular device. Hardware anecdotes do not establish a universal requirement.
There is no hosted token price assigned to this downloadable model. Measure generation time, retries and accepted assets on the runtime you intend to operate. Keep the original prompt, seed, dimensions and settings with the output so a later revision can be compared fairly.
Compare ERNIE Image and ERNIE Image Turbo for the same brief, or consider Qwen Image 2.1 when image editing and transparent output are central to the workflow.
Publisher source 1 Deployment limits and commercial terms belong to the selected checkpoint or API route.
Compare the model on the work you need to complete.
ERNIE Image Turbo is listed for Text-to-image. It accepts text and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
Choose the output size and quality your design needs. Include drafts and revisions when budgeting an accepted image.
The weights are downloadable under Apache-2.0. Check runtime support and size the complete model and working context for your expected concurrency.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
Alibaba
Qwen-Image-2.1-Turbo creates and edits images, including reference-guided and transparency workflows; hosted billing is separate from local deployment.
Alibaba
Qwen-Image-2.1-Pro creates and edits images, including reference-guided and transparency workflows; hosted billing is separate from local deployment.
Meta
Muse Image 1.0 generates and edits images through conversational or single-shot endpoints, with flat per-image Meta Model API billing.
krea
Krea 2 Raw: a base checkpoint for further training, not the recommended everyday inference release.
krea
Krea 2 Turbo: the post-trained and distilled Krea 2 checkpoint for image generation.
OpenAI
GPT-5.4 Image 2 from OpenAI: Image model for prompt-driven generation, editing, and visual design workflows.