Skip to main content
Wsong.ai

Wsong.ai model guide

Model V4.5 vs V5: what actually changes

Wsong.ai currently uses two provider execution paths. Free accounts start on V4.5; the selector presents V5 to active paid subscribers. The evidenced differences are access, route, and track length—not a promise that every model label has a different sound.

Last reviewed:

Start writing free

Model V4.5

Free default

Up to 4 minutes

V4.5 execution path

The default model for free accounts in the Wsong.ai selector. It is the practical starting point when a track can stay within four minutes.

Model V5

Paid option

Up to 8 minutes

V5 execution path

The model shown to active paid subscribers in the Wsong.ai selector. It uses the V5 provider route and supports tracks up to eight minutes.

Compatibility labels

What Model V1 and Model V3 mean on Wsong.ai

Those labels remain available for compatibility and reference. Under the current generation contract, V1, V3, and V4.5 requests all execute on the V4.5 provider route. We therefore do not assign V1 or V3 a separate quality, style, or capability promise.

Why this distinction matters

A visible selector label is not evidence of a separate backend model. For a reproducible comparison, use V4.5 and V5—the two routes that actually change the provider request.

Two execution paths

V1, V3, and V4.5 requests use the V4.5 route. Only a V5 selection changes the provider request to the V5 route.

A known length boundary

Model V4.5 supports tracks up to 4 minutes. Model V5 raises that documented limit to 8 minutes.

Access is not usage rights

Model access, download access, and commercial usage rights are separate product rules. Check Pricing and the License Agreement for the latter two.

Run a fair V4.5 vs V5 comparison

  1. 01

    Keep the input fixed

    Use the same prompt, lyrics, style, title, vocal setting, and instrumental setting for both models.

  2. 02

    Respect the length boundary

    Choose V5 when the intended song may need more than four minutes; otherwise length alone does not require it.

  3. 03

    Name the decision

    Judge a specific target such as lyric clarity, vocal phrasing, section contrast, or ending—not a vague sense of quality.

The current real-song gallery compares two outputs from the same selected route; it is not a controlled cross-model test. Hear the four paired generation examples.

V4.5 and V5 questions

What is the difference between Model V4.5 and V5?

They use different provider execution routes. Model V4.5 is the free-account default and supports tracks up to 4 minutes. Model V5 is the paid selector option and supports tracks up to 8 minutes. Wsong.ai does not claim that either route will always produce the better song.

Can free users write a song with Model V4.5?

Yes. After sign-in, free accounts start with Model V4.5. The Free Plan currently includes 3 generations—up to 6 public songs a day—for online listening.

Are Model V1 and V3 separate execution models?

The V1 and V3 labels are still accepted for compatibility, but current generation routing sends V1, V3, and V4.5 requests through the same V4.5 provider path. They are not documented here as three independently evidenced model backends.

Who can select Model V5?

The current selector presents Model V5 to active paid subscribers and redirects free users to pricing. Generation limits and downloads are defined by the Pricing page; commercial usage rights are a separate question defined by the License Agreement.

Does each generation still create two songs?

Yes. Each Wsong.ai generation returns two versions from the same input and selected model. Keep the prompt, lyrics, style, and model fixed if you want to judge the two interpretations fairly.

Start on V4.5, move to V5 for a reason

Free accounts can test the full writing workflow on V4.5. Choose V5 from an active paid account when you need its execution route or the longer track limit.