xAI recommends Grok Imagine Video 1.5 ($0.08/s) over the $0.05 model

xAI lists grok-imagine-video-1.5 at $0.080 per second and grok-imagine-video at $0.050 per second, and its docs recommend Grok Imagine Video 1.5. So the 5 cent model is the cheaper one, but it is not the one xAI points new work to. Sume's catalog carries only grok-imagine-video-1.5, as an image-to-video model.

The two xAI models

Prices are from the xAI models page. The totals are my arithmetic, rate x seconds.

When the cheaper model still fits

1.5 costs 60% more per second ($0.080 / $0.050 = 1.6). The page I read does not say how the two models differ in output, so the reason to pay the premium is whatever you see in your own test. The xAI guide says that duration and resolution both affect the total cost, and that video can be up to 15 seconds.

How to decide

Pick by what you can verify. Start from xAI's recommendation, 1.5, and test the cheaper model only if the cost gap matters at your volume. At 1,000 clips of 10 seconds, the gap is 1,000 x 10 x ($0.080 - $0.050) = $300 at xAI's listed rates. If the cheaper model's output meets your bar, that is real money. If it does not, 1.5 is the model xAI points you to.

• Run the same 5 prompts through both models and compare side by side.
• Check the length: the xAI guide says video can run up to 15 seconds, and it names only 1.5 as the model id.
• On Sume, the choice is already made: only 1.5 is listed.

What Sume lists

Sume's Video Router lists grok-imagine-video-1.5 at 480p and 720p, for 4 to 15 seconds, as image-to-video only. It has no text-to-video mode and no reference input on Sume. Sume lists no grok-imagine-video. The Sume price per second is on GET /v1/videos/models, and I do not state it here.

xAI's guide says video generation starts from one source still image. That matches the Sume row.