Module 22.2 closed on still vendors. Volume 22.3 is Video AI as an ecosystem catalog: who you buy when motion is the job. Runway (Runway ML) is the filmmaker-workstation pole—text-to-video, image-to-video, video-to-video, motion brush / camera controls, Director-style tools, and a documented API. Runway sold “AI for editors” years before Sora’s demo.
Volume 16.4 already cataloged Runway Gen-3 vs Sora / SVD. Pin Gen-3 vs later Gen-4 / Aleph-class SKUs in every eval. Volume 17 (diffusion, temporal layers on SVD) explains why I2V is easier than long T2V—Runway stays closed SaaS. A 22.2 still (FLUX, Firefly, MJ) often becomes the I2V start frame. Next: Pika for creator/effects culture.
Learning Objectives
By the end of this lesson, students should be able to:
- Place Runway as a closed video vendor with T2V, I2V, V2V, and director controls plus an API.
- Score Pricing / Strengths / Weaknesses qualitatively vs Kling, Veo, Pika, and open SVD.
- Choose Turbo vs higher-quality SKUs as a latency/credit trade-off (no fake $).
- Sketch the Runway API async job pattern (model string + prompt + assets + poll).
- Cross-link Vol. 16.4 Gen-3 catalog and Vol. 17 diffusion / SVD for why I2V locks better.
- Know when Kling (motion), Veo (GCP), Pika (effects), or SVD (open) is the better aisle.
Runway is Runway ML’s generative video company/product. The 2024 workhorse family was Gen-3 (Alpha / Turbo); later Gen-4 / Aleph-class editors extend the line. It generates short clips from text, from a start frame (I2V), or by transforming an input clip (V2V / style/motion controls). Product surfaces include the Runway web app, timeline-adjacent tools, and the Runway API. Weights are not published. Pin the model string in every spike.
Pricing, Strengths & Weaknesses
| Axis | Runway | Kling / Hailuo | Veo / open SVD |
|---|---|---|---|
| Pricing | Seat plans plus credit meters (Turbo burns fewer credits than higher-quality SKUs). API billed on jobs/duration. No fake dollar table—read current credit sheet. | Regional credit packs; often spiked as “value,” still a meter. | Veo: GCP / Gemini quotas. SVD: your GPU hours + Stability license. |
| Strengths | Director controls (I2V, V2V, camera, motion brush, keyframes—feature set evolves); mature API; editor-native UX; I2V from a 22.2 still. | Motion-forward look-dev (Kling); credit-value spikes (Hailuo)—measure. | Veo: GCP + SynthID + audio SKUs. SVD: open short I2V, Vol. 17 learning. |
| Weaknesses | Closed; credit burn on iteration; Turbo trades quality; not realtime 24 fps edit; audio often still NLE; Gen-3 vs Gen-4 mix-ups ruin A/Bs. | Weaker editor/API story vs Runway; region/ToS gates. | Veo: fewer brushes. SVD: not T2V cinema. |
Modes Filmmakers Actually Use
Text-to-video
- Look-dev / B-roll ideas
- Weaker than I2V for brand lock
- Prompt like a shot list
Image-to-video
- Lock a still from 22.2 / 16.3
- Animate camera + subtle motion
- Common production path
Video-to-video / tools
- Restyle existing plates
- Motion brush / regions
- Why editors pick Runway
API Pattern (Pin the Model String)
Runway’s API uses named models (historically strings like gen3a_turbo—confirm current enum). Jobs are async. Treat Turbo as the latency SKU, not a secret quality upgrade. Vol. 17: the engine is still iterative denoising; the product is credits + tools.
When Runway Is the Right Vendor
Pick Runway when
- Editors need V2V and regional motion
- I2V from a locked 22.2 still
- Studio already pays Runway credits
- You want a vendor API, not Discord-only
“Gen-3 Turbo is just Gen-3 with a marketing name; quality is identical, and Runway is realtime like a game engine.” Turbo exists to burn fewer credits and return faster—expect a quality/coherence trade-off. Jobs are async, not interactive 24 fps. Do not cite a Gen-4 demo as a Gen-3 eval. Pin model, duration, and whether the job was T2V, I2V, or V2V. Vol. 16.4 Gen-3 is the SKU catalog; this page is the vendor scorecard; Vol. 17 is why I2V denoising locks a still better than prose T2V.
Knowledge Check
- Short Answer: What is Runway’s catalog differentiator vs Sora/Veo prompt cinema? Answer: Filmmaker/director controls (I2V, V2V, camera/motion tools) + a mature API.
- True/False: Runway weights are open like SVD. Answer: False—closed Runway SaaS.
- Multiple Choice: Best mode to lock a 22.2 brand still: (a) pure T2V, (b) I2V, (c) OCR. Answer: (b).
- Short Answer: Why does Turbo exist (pricing/ops, no $)? Answer: Lower latency / credit cost with a likely quality trade-off.
- True/False: You should mix Gen-3 and Gen-4 results in one unlabeled A/B. Answer: False—pin model strings separately.
- Multiple Choice: V2V is primarily: (a) restyling/transforming an input clip, (b) STT, (c) RAIL-M. Answer: (a).
- Short Answer: Which Vol. 16.4 lecture is the Gen-3 catalog sibling? Answer: Runway Gen-3.
- True/False: Runway latency is typically realtime like a game engine. Answer: False—async jobs; Turbo is faster, not interactive 24 fps edit.
- Multiple Choice: Next 22.3 vendor (creator/effects): (a) Pika, (b) Canva, (c) LangGraph. Answer: (a).
- Short Answer: When is SVD still relevant if Runway exists? Answer: Open weights, on-prem I2V, Vol. 17 learning, no SaaS ToS.
Key Takeaways
- Runway is the closed director-oriented video vendor—tools + API, credit-metered.
- I2V and V2V often beat pure T2V for production lock; Turbo vs Alpha is a credit trade.
- Pin model string and mode; do not mix Gen-3/Gen-4 unlabeled.
- Vol. 16.4 catalogs Gen-3; Vol. 17/SVD explain I2V denoising.
- Continue with Pika.
Lab: Same 22.2 still: Runway I2V (Turbo if credits are tight) vs T2V-only. Score brand lock and camera smoothness. Record credit burn if visible—no fake FPS leaderboard.
Whiteboard: Shot pipeline: 22.2 still → Runway I2V → NLE + real type + audio. Contrast “one Sora prompt = finished film.” Arrow to Vol. 17 diffusion.
Recap: Runway is closed video with filmmaker controls and an async API—pin SKU and mode. Continue with Pika.