Kling vs Runway: models, clip limits and API prices
Kling's API sells only Kling models; Runway's sells Gen-4.5 and Aleph 2.0 plus other labs' models. Clip limits, audio and per-second prices compared.

Kling and Runway both make AI video models and both sell API access, but they sell different catalogs. Kling's API sells only Kling models, led by Kling 3.0 (3–15 second clips, up to 4K, optional native audio), billed per second in Units drawn from resource packages. Runway's API sells its own Gen-4.5 (2–10 second clips) and Aleph 2.0 next to other labs' models, billed in $0.01 credits, and no Kling model appears in its model list.
Every Kling and Runway fact below comes from that company's own developer docs, read on 2026-09-28 and listed under Sources. Which one makes the better-looking clip is not answered here: run the same prompt on both. The Sume section uses Video generation.
How do Kling and Runway compare?
| Item | Kling | Runway |
|---|---|---|
| Models on the API | Kling only: 3.0 Turbo, 3.0, 3.0 Omni, O1, 2.6, 2.5 Turbo | Its own Gen-4.5, Aleph 2.0, Gen-4 Turbo and Act-Two, plus third-party models such as Wan, Grok Imagine and Gemini Omni Flash |
| Flagship clip length | Kling 3.0: 3–15 s | Gen-4.5: 2–10 s |
| Flagship inputs | Kling 3.0: text, image or video | Gen-4.5: text or image |
| Flagship resolution | Kling 3.0: 720P, 1080P or 4K | Gen-4.5 text to video: 1280:720 or 720:1280 |
| Native audio | Kling 3.0: settings.audio is native or off (default off) | Per model: seedance2 and veo3.1 take an audio flag, default true |
| Billing unit | Units per second, deducted from resource packages | Credits per second, bought at $0.01 each |
| Auth | Authorization: Bearer with an API key | Authorization: Bearer plus an X-Runway-Version header |
Can Runway's API run Kling models?
Not today. Runway's models page lists wan3, seedance2_5, grok_imagine_1_5, seedance2, h3_max, aleph2, gen4.5, gen4_turbo, act_two, veo3.1, gemini_omni_flash and a few more, and no Kling id. The reverse holds too: Kling's capability map lists only Kling models. So "Kling vs Runway" is also a choice between one lab's lineup and a catalog that mixes labs.
Kling vs Runway pricing: what does a second cost?
Kling bills per second from resource packages, with the price set by model, resolution and audio. Runway bills credits per second per model, and a generation routed through a Model Router is billed at the rate of the model it picks (Runway pricing). A 5-second clip is 3 Units ($0.42) on Kling 3.0 at 720P without audio, and 60 credits ($0.60) on Gen-4.5.
| Model | Price per second |
|---|---|
| Kling 3.0, no native audio, 720P | 0.6 Units ($0.084) |
| Kling 3.0, no native audio, 1080P | 0.8 Units ($0.112) |
| Kling 3.0, native audio, 720P | 0.9 Units ($0.126) |
| Kling 3.0, 4K | 3.0 Units ($0.42) |
Runway gen4.5 | 12 credits ($0.12) |
Runway aleph2 | 28 credits ($0.28), 56-credit minimum |
How do the two APIs differ to call?
Both take a request, return a task id, and finish asynchronously. Kling 3.0 is POST /text-to-video/kling-3.0 with prompt, settings (duration 3–15, resolution, aspect_ratio 16:9, 9:16 or 1:1, audio) and an optional options.callback_url (Kling 3.0 text to video). Runway is POST /v1/text_to_video with a body keyed on model; you poll GET /v1/tasks/{id} until the status is terminal, and output URLs are temporary, so download what you keep (Runway Dev context). Does Runway have an API? covers Runway's keys and SDKs.
Can I call Kling 3.0 without a Kling account?
Yes, through a multi-model API. On Sume today, Kling 3.0 is kling-3 on POST /v1/videos, billed at provider list × 1.25 (Video generation): $0.14 a second silent or $0.21 with audio, plus a 5.5% agent fee by default (API pricing). Runway's Gen and Aleph models are not in Sume's catalog. Kling 3.0 API covers the kling-3 limits, and Sume vs Kling compares Kling's own billing with Sume's.
Which should I pick?
- Pick Kling if you want Kling models only, clips up to 15 seconds, 4K, or its per-second Unit prices.
- Pick Runway if you want its own Gen-4.5 or Aleph 2.0, or several labs' models on one Runway account.
- Pick a multi-model API if Kling 3.0 is one of several models your code calls on one balance.
- Whichever you pick, compare the same resolution and audio setting: every price above changes with both.
Sources
- Kling API: Video capability map (read 2026-09-28)
- Kling API: Video pricing (read 2026-09-28)
- Kling API: Kling 3.0 text to video (read 2026-09-28)
- Kling API: Authentication (read 2026-09-28)
- Runway: Available AI models (read 2026-09-28)
- Runway: API pricing (read 2026-09-28)
- Runway: API reference (Markdown) (read 2026-09-28)
- Runway Dev context (read 2026-09-28)
- Runway Dev: complete documentation (read 2026-09-28)
- Video generation
- API pricing
Related posts
More in Models
- Lip sync AI API options: photo or video input, limits, price
Lip sync APIs take a face and audio and return a talking clip. Some animate a photo, others redub an existing video. Inputs, limits and price units.
- Lyria MCP: generate music with Lyria 3.5 from Claude
Sume's hosted MCP server has a music_create tool that runs on Google Lyria 3.5 today: a prompt in, a job id back, an audio file on media.sume.com.
- Multi-shot video generation: what it is and how to do it
Multi-shot video generation returns one clip with several cuts. Which APIs offer it (Kling 3.0, Runway) and how to build it shot by shot on Sume.
- Nano Banana MCP server for Claude: how to add one
To use Nano Banana in Claude, connect an MCP server that runs it. Which servers list it, how to add Sume's, and what a call looks like and costs.
Written by Sume