AI video news Oct 1-8, 2026: Vidu Q4 Preview, Kling 4.0, Sume coverage
Vidu Q4 Preview launched Oct 7 and kling.ai headlines Kling 4.0. We checked Runway, Luma and Veo, and which of them Sume lists.

The one dated AI video launch we could confirm from a primary page this week is Vidu Q4 Preview, released on 2026-10-07 with up to 15 image references, up to 3 audio references and output up to 4K. Kling AI's home page now headlines Kling 4.0 as its newest model, though the page text we read gives no date. The Runway, Luma and Google pages we read show nothing newer than September 3, June 9 and Veo 3.1 respectively. Sume lists none of Vidu Q4 Preview, Kling 4.0, Runway, Luma or Veo in its video catalog.
What the vendor pages say
Vidu's launch release is dated October 7 and says availability is through its web product and API. The Kling home page lists Kling 4.0 first, then Kling Image 3.0 and Kling Video 3.0. Runway's research page lists GWM Worlds 2 on September 3 and Solaris on September 1. Luma's news page lists Ray 3.2 on June 9. Google's Veo page names Veo 3.1 as the latest version.
| Model or product | What the page read shows | In Sume's video catalog? |
|---|---|---|
| Vidu Q4 Preview | Released Oct 7, 2026; web and API | No |
| Kling 4.0 | Headlined on kling.ai; no date in the text read | No; kling-3 is listed |
| Runway GWM Worlds 2 | Research preview dated Sept 3 | No |
| Luma Ray 3.2 | Dated June 9 | No |
| Veo 3.1 | Latest version on Google's page | No; not in the v1 catalog |
| MiniMax H3 | Announced July 31, 2026 | Yes: minimax-h3 and minimax-h3-max |
What you can generate on Sume today
Sume's video catalog is the list returned by GET /v1/videos/models: seedance-2.5, seedance-2-mini, seedance-2, seedance-2-fast, kling-3, wan-3.0, grok-imagine-video-1.5, minimax-h3, minimax-h3-max, gemini-omni-flash-1.1 and h3-max-recast, plus higgsfield-genjutsu where its provider is configured. If a brief calls for a model that is not on this list, the closest substitutes are in the posts linked from this article.
Steps to test a new model's brief on Sume:
- Fetch the catalog and write down the row's limits for references, duration and resolution.
- Choose a row that covers your must-haves; drop the nice-to-haves.
- Run a 5 second draft at a low resolution before a full render.
- Compare the draft against the target model's own output if you have access.
curl "https://api.sume.com/v1/videos/models" \
-H "Authorization: Bearer $SUME_API_KEY"How to read a launch headline
A launch post tells you what a vendor will sell; it does not tell you what an aggregator can call. Three questions sort most headlines. Is there a public API, or only a web product? Does the page give a date? Is the model id one you can send in a request body? For Vidu Q4 Preview the first and third answers exist on the vendor's developer pages, with the model named viduq4-preview. For Kling 4.0 the page we read only headlines the name.
None of this means a model is better or worse than what Sume lists. It means you should compare on the limits you need, such as duration, references and resolution, rather than on the name.
What Sume does not do
Sume does not claim a launch is available until the catalog lists it, and does not rename another model to match a headline. We also did not verify any benchmark or ranking claims from outside the vendors' own pages, so none appear here. Check the list on the day you build.
Sources
- PR Newswire: Vidu Launches Q4 Preview (read 2026-10-08)
- GlobeNewswire: Shengshu Technology Launches Vidu Q4 Preview (read 2026-10-08)
- Kling AI home page (read 2026-10-08)
- Runway research page (read 2026-10-08)
- Luma AI news (read 2026-10-08)
- Google DeepMind: Veo (read 2026-10-08)
- Sume docs: Video generation
- Sume docs: Video Router
Related posts
More in Models
- Audio API launches, Oct 1-8 2026: dates, prices and what Sume covers
A dated calendar of speech, voice and music releases from Sep 28 to Oct 8, 2026, each with the number the vendor published, and which ones Sume lists.
- background: transparent on the Sume image API: two GPT rows only
Transparent PNGs from the Sume image API need background: transparent, which only the two GPT Image 2.5 rows list. See the request, price and the 400 elsewhere.
- Best image-to-video model, October 2026: Wan 3.0 is 7th at 1,164 Elo
On the AA image-to-video board MiniMax H3 Max leads at 1,195 and Wan 3.0 is seventh at 1,164. Which of the top 12 you can call on Sume, with per-minute prices.
- Can I show HunyuanVideo clips to EU viewers? Section 5(c) read
Tencent's HunyuanVideo license says do not use or display Outputs outside the Territory, which excludes the EU, UK and South Korea. Clause by clause.
Written by Sume