April 2026 Updates: WAN 2.7 pricing, simpler API integrations, and more format support

April was all about two things: better model economics, and a cleaner developer experience.

We added a new default video model in AI Hub with a launch price that makes experimentation much easier. We also shipped platform updates that remove integration edge cases across Creative APIs, plus expanded format support and more control over GenAI video outputs.

AI Hub updates: WAN 2.7 becomes the default

Alibaba WAN 2.7 is live with launch pricing

A strong default model matters. It sets the baseline for quality, speed, and cost for every team that tries video generation for the first time.

This month, we made Alibaba WAN 2.7 the default video model in Picsart AI Hub, and we launched it with pricing that is designed to be used, not just admired.

What’s new:

• WAN 2.7 is now available in AI Hub for video generation workflows.
• Launch pricing is 10 credits per second.
• Many other video models are 50 credits per second, so this is a major cost step-down for testing and production.

If you want the short version: you can start with a strong default now, then switch models later without changing your integration.

Developer platform updates: more consistency across Creative APIs

A lot of teams told us the same thing in different ways: “We love the capabilities, but integrations should feel more uniform.”

So we shipped a set of updates that move Creative APIs toward one standard approach, across endpoints and product families.

Standard sync and async control with the Prefer header

We are standardizing how clients control sync vs async behavior using the HTTP Prefer header.

What this unlocks:

• A consistent pattern across services
• Cleaner fallbacks for long-running jobs
• Less endpoint-by-endpoint logic in your integration

How it works:

• Requests are synchronous by default for most endpoints.
• You can force async behavior with Prefer: respond-async.
• You can also request “best effort sync” with Prefer: wait=<seconds>. The gateway waits up to your limit, then falls back to async with a 202 if it is not ready.

If you previously used a mode parameter in some services, note that mode is deprecated and Prefer is the forward path.

For more details, read the Sync and Async Execution docs.

Deprecation deadline: _id input parameters will be removed on June 1, 2026

This is an important cleanup and it impacts real integrations.

We are deprecating input parameters that end in _id across Programmable Image APIs and GenAI APIs. Examples include:

• image_id
• reference_image_id
• mask_id
• other similar *_id inputs

Deadline:

• Support for _id inputs will be removed on June 1, 2026.

What to do:

• Audit your usage for any *_id parameters.
• Move to *_url inputs or binary upload inputs where supported.

GenAI status values are now standardized

We are standardizing GenAI job statuses to remove “almost the same” states that made polling logic harder than it should be.

Moving forward, GenAI job status values are:

• success
• error
• processing

Best practice:

• Write your switch logic around success and error.
• Treat anything else as “keep waiting”, which is now simply processing.

More request options: JSON, multipart, and form URL-encoded

Creative APIs now support a consistent set of request formats so you can integrate in a way that fits your stack.

Supported request content types:

• application/json
• multipart/form-data
• application/x-www-form-urlencoded (form URL-encoded)

Guidance:

• Use JSON when you provide inputs as URLs and you want the simplest request construction.
• Use multipart when an endpoint supports binary image upload and you need to inline a local file.
• Use form URL-encoded when your client or automation tool outputs it naturally.

Base64 support:

• For image-based endpoints that accept image_url, you can also provide a base64 data URI in image_url if you prefer not to host the file.

Read more about Request content types

Format support: BMP is now accepted for image inputs

BMP support added for image inputs

If your pipeline includes BMP files, you no longer need a pre-conversion step before calling Picsart image endpoints.

Our supported image input formats include:

• JPG, PNG, WEBP, TIFF
• BMP
• plus additional formats such as MPO, MJPEG, and HEIC (input only)

Read more about Supported input and output formats

GenAI video generation: new export parameter for output control

More control over video outputs with export

GenAI video generation services now support an export parameter, similar to Programmable Video APIs.

This gives you more granular control over output settings such as:

• FPS selection
• codec and container options
• other export-related parameters used in production workflows

If you build automated pipelines (ads, variations, batch generation), export-level control is one of the fastest ways to reduce manual post-processing.

GenAI terms update: new enterprise terms URL

Generative AI Additional Terms updated for enterprise customers

As of April 30, 2026, our Generative AI Additional Terms for enterprise customers are available at a new URL:

• https://picsart.com/genaitermsofuse-enterprise/

If you use Picsart GenAI services, we recommend reviewing the updated terms and sharing them with your legal and compliance stakeholders.

Bug fixes and reliability

We also shipped a set of bug fixes and stability improvements across Creative APIs and GenAI services. These updates focus on reliability, response consistency, and smoother integration behavior in real production conditions.

Looking ahead

April was a “platform foundations” month. Better defaults, fewer integration branches, and more predictable behavior.

If you have feedback or want to sanity-check a migration plan (Prefer header, _id removal, or request formats), contact our team.

Looking for the full change log?
👉 Browse the API release notes