Runway and Luma AI have moved AI video generation closer to everyday software by releasing APIs for their video models within hours of each other. The launches give companies and developers a more direct way to bring generated video into their own apps and services.
The two releases also sharpen the competitive picture around AI video. Instead of keeping video generation mainly inside standalone tools, both companies are making their systems available as building blocks for other products.
What Runway Released
New York-based Runway has launched an API for its Gen-3 Alpha Turbo AI video model. The company is not opening access to everyone at once. Its API is being rolled out gradually, beginning with select partners, and interested parties can join a waitlist.
Runway says advertising group Omnicom is already using the API. The company did not provide details about the specific application, so the clearest takeaway is that the API is already being used by at least one major business customer.
Runway's pricing starts at one cent per credit. The source article states that five credits are needed for a one-second video, which gives developers a simple starting point for thinking about cost when testing short AI video outputs.
What Luma AI Released
San Francisco-based Luma AI followed with an API for its Dream Machine model. Unlike Runway's gradual rollout, Luma AI's API is available immediately.
Luma AI charges $0.32 per million pixels generated. According to the source, that works out to about $0.35 for a five-second 720p video.
The Dream Machine API includes several capabilities that matter for app builders. It supports text-to-video generation, image-to-video conversion, and camera motion control. Together, those features give developers more than one path for creating video: they can begin from written prompts, from still images, or from controlled camera movement.
Why APIs Matter for AI Video Generation
An API changes how AI video generation can be used. A standalone model can be impressive, but an API lets another product call that model directly. That means video generation can become part of a larger workflow, service, or application experience.
For developers, this can reduce the gap between experimenting with a model and offering AI video features to users. For businesses, it creates a way to connect generated video to existing products without building the underlying model themselves.
The source article describes both APIs as tools for incorporating AI video generation into apps and services. That framing is important because it points to a shift from demo-like usage toward integration. The model becomes infrastructure that other software can depend on.
Pricing and Access Are Different
The two launches are similar in purpose but different in availability and pricing structure. Runway is beginning with select partners and a waitlist, while Luma AI is available immediately.
- Runway: API for Gen-3 Alpha Turbo, gradual rollout, select partners first, waitlist available, pricing starts at one cent per credit, five credits needed for a one-second video.
- Luma AI: API for Dream Machine, available immediately, priced at $0.32 per million pixels generated, about $0.35 for a five-second 720p video.
Those differences matter for teams evaluating which API they can test now. Runway may be available first through a more controlled partner process. Luma AI is positioned for immediate access, with cost tied to the amount of video generated in pixels.
Competition and Responsible Use
The releases intensify competition in AI video generation. The source article notes that Adobe recently introduced its "enterprise-grade" Firefly Video AI model, but it is not yet available as an API. OpenAI also has a powerful video model called Sora, but has held it back so far.
That leaves Runway and Luma AI pushing API access forward while other major names remain at different stages of availability. For developers and businesses, the practical issue is not only which model is powerful, but which model can actually be integrated into a product.
Both companies also emphasize responsible use of their technology. Luma AI says it uses a multi-level moderation system that combines AI filters with human oversight. The source does not provide further detail on Runway's safety approach, beyond noting that both companies stress responsible use.
The result is a meaningful step for AI video generation: two startups are giving developers programmatic access to video models, with different rollout strategies, different pricing models, and a shared focus on making the technology usable inside other applications.