Kling O3 4K on Runway MCP: where 4K video comes from on Sume
Runway's Sept 18 changelog added Kling O3 4K to its MCP. Sume lists no Kling 4K row; its 4K options are Gemini Omni Flash 1.1 and priced MiniMax H3 upscales.

Runway's changelog says Kling O3 4K, Kling 3.0 Standard, Kling O3 Standard and Kling 3.0 Motion Control became available on Runway MCP on 2026-09-18. Sume does not list a Kling 4K row. The docs name Kling only as a Motion Control tool in the hosted MCP list. If 4K is the requirement, Sume's options are gemini-omni-flash-1.1 at 4K and minimax-h3 at a priced 4K upscale.
So the honest answer to "can I get Kling O3 4K through Sume" is no; the useful answer is which 4K routes Sume does have.
Three routes to 4K compared
The table lists only what the vendor pages and Sume's catalog state. It does not rank output quality, which you should judge on your own footage.
| Route | 4K | Length | Notes |
|---|---|---|---|
| Runway MCP: Kling O3 4K | yes, per the changelog entry | not stated on the entry | Runway credits apply; sign-in through Runway |
Sume gemini-omni-flash-1.1 | yes, 4K generation | 3 to 10 seconds | 16:9 or 9:16, native synced audio always on |
Sume minimax-h3 | 2K and 4K are upscales of native 768p, priced if requested | 5 to 15 seconds | minimax-h3-max has no 2K or 4K |
What 4K costs on Sume
Sume bills the provider list times 1.25, as the Video Router page states. In Sume's catalog the Gemini Omni 4K list is $0.30 per second, which works out to about $0.375 per second billed, so a 10-second 4K clip is about $3.75. The H3 4K upscale list is $0.16 per second, about $0.20 billed, so 10 seconds is about $2.00. Read the live rate first.
- Gemini Omni 4K, 10 seconds: about $3.75.
- MiniMax H3 4K upscale, 10 seconds: about $2.00.
- Both numbers are estimates from the catalog rate card, not quotes.
If you want the Kling look
Sume's docs list Kling as a Motion Control tool, kling-motion-control_create, in the hosted MCP tool list, and give no 4K Kling row. For 4K from Kling itself, use Runway's MCP or Kling's developer platform, whose model page lists 4K, multi-reference, motion control and audio as features of Kling 3.0 and O3.
A mixed workflow is possible: draft on a cheap Sume row, then re-render the shot you approve on the 4K route you chose. Keep the first frame or a reference image the same.
A 4K request on Sume
This asks gemini-omni-flash-1.1 for 4K. The row only accepts 16:9 or 9:16.
curl -X POST https://api.sume.com/v1/video-router/generate \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: omni-4k-001" \
-d '{
"model": "gemini-omni-flash-1.1",
"prompt": "A glass perfume bottle on wet stone, slow orbit, rain sound",
"resolution": "4K",
"duration": 6,
"aspect_ratio": "16:9",
"mode": "async"
}'Source notes
The Runway entry is on Runway's changelog. Check Video generation for the response shape and the API reference for errors.
Sources
Related posts
More in Comparisons
- Lip Forcing video lip-sync: 37 GB GPU or a hosted API?
Lip Forcing is an open video-to-video lip-sync model that needs about 37 GB of GPU memory. What it does, what it asks of you, and the hosted alternatives.
- Live avatar agent or rendered video: D-ID vs Sume
D-ID V4 Expressive Visual Agents are live, LLM-connected avatars. Sume avatar video is rendered from a script in 4 to 60 s. Which one fits which job.
- How long can an AI song be? Suno, ElevenLabs and Lyria compared
ElevenLabs Music runs 3 s to 5 minutes; Suno v6 is reported at up to 8 minutes; Lyria 3.5 makes a couple of minutes. Join tracks with a $0.01 concat.
- LTX-2.5 license ($10M ARR line) vs hosted video ids
LTX-2.x Community license is free under $10M ARR and paid above it. Here is what changes if you call a hosted Sume video id instead of running weights.
Written by Sume