Luma Ray3.2 API: the docs page still lists ray-2 and ray-flash-2
Luma says the Ray3.2 API is live. The docs page I read on Oct 7 lists only ray-flash-2 and ray-2. How to check before you code, and Sume's gap.

Luma's Ray3.2 announcement, dated June 9, 2026, says the Ray3.2 API is available for the first time. The video generation docs page I read on October 7 lists only two models, ray-flash-2 and ray-2. Before you code to a Ray3.2 id, confirm it in Luma's current API reference, since that page did not show one. Luma is not in the Sume catalog.
What each Luma page says
| Page | What it states |
|---|---|
| News, June 9, 2026 | Ray3.2 up to 20 s at 1080p, native HDR, 16-bit EXR, up to 16 keyframes, API available for the first time |
| Docs, video generation | Models ray-flash-2 and ray-2; resolutions 540p, 720p, 1080p, 4K; loop, keyframes, concepts, callback_url |
Why the gap may exist
Docs pages and announcements often update on different schedules, and a model can be live in one listing and missing from another. I did not find a Ray3.2 model id on the docs page, and I do not state one here. If you are writing integration code, copy the id from the response of Luma's own models listing or reference page, not from a blog post.
What Sume offers meanwhile
Sume cannot call Luma. For a long 1080p clip, wan-3.0 reaches 30 seconds and seedance-2.5 also reaches 30, in one request each. Neither does native HDR or EXR output, which Luma lists for Ray3.2. If HDR is the reason you want Ray3.2, plan on Luma's API.
A checklist before integrating
- Open Luma's current API reference, not a cached copy, and look for the Ray3.2 model string.
- Send a single cheap request and read the response model field back.
- Check that keyframes, loop and callback_url still work on the Ray3.2 id; the page I read lists them on ray-2.
- Keep your resolution and length requests inside what the announcement states: 20 seconds at 1080p.
What this does not claim
I do not say the Ray3.2 API is missing, and I do not say it is live under any particular id. The two pages I read disagree by omission, and the news post is a month older than a page that may have been updated since.
Sources
Related posts
More in Developers
- lyria-3-pro-preview vs Sume lyria-3-pro: which model string?
Google lists lyria-3-pro-preview at $0.08 per song. On Sume the Music Router id is lyria-3-pro, and a Google string fails with model_not_found.
- Mandarin Chinese speech to text API: Sume STT language_code zh
Transcribe Mandarin audio with Sume STT using language_code zh, then check the result and timings. $0.01 per audio minute and no accuracy claim without a test.
- mask_url on ChatGPT Image 2: not listed, only the 2.5 rows take it
Sume lists mask_url only on ChatGPT Image 2.5 Flare and Sunburst. Sending it to ChatGPT Image 2 is rejected. What to do for a masked edit on the older row.
- MCP outputSchema vs Sume output_schema: who sets the contract
In MCP the server declares a tool's outputSchema. In a Sume Agent Completion you send output_schema per run, and the result can still come back degraded.
Written by Sume