Luma AI
Directable AI video with Ray3.2, inside a multi-model creative agent
Overview
Luma AI is a Palo Alto company whose Luma Agents platform plans creative work and generates video, images and audio with its own Ray and Uni models and with third-party models. Its current video model, Ray3.2, supports up to 16 keyframes per clip, HDR with EXR export and clips of up to 20 seconds, and is also sold as an API. Its older 3D Capture iPhone app is still available.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 14 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- 10,000 credits
- Luma and third-party image and video models
- Edit access for guest collaborators
- Commercial use
- Everything in Plus
- 40,000 credits
- 4x usage with the Luma Agents
- Everything in Pro
- 150,000 credits
- 15x usage with the Luma Agents
- Manage and add team members
- Projects, team organisation and team-wide sharing
- Usage analytics and shared team credits
- SSO
- Everything in Team
- Enterprise commitments
- Dedicated education and training
- Custom fine-tuning
Use cases
What you can produce with Luma AI
- Text-to-video and image-to-video clips of up to 20 seconds at 1080p with Ray3.2
- Shots directed frame by frame with up to 16 keyframes in a single clip
- HDR video with 16-bit EXR export for grading and compositing pipelines
- Existing footage restyled or reframed with video-to-video and Reframe
- A static ad resized to five placements and translated into several languages with Variants
- Multi-step creative projects planned and generated by Luma Agents across Luma and third-party models
- Ray3.2 video generation built into your own product through the Luma API
Luma is a serious option for agencies and studios that want directable video: Ray3.2 adds frame-level keyframes, HDR with EXR export and 20-second clips, and Luma Agents put Luma's own models alongside Veo, Kling, Seedance and others in one workspace. A $900 million Series C led by Humain means the product is well funded.
The catch is cost control. There is no free plan, credits do not roll over, and Trustpilot reviewers repeatedly describe failed generations draining a month's allowance. Start on Plus, watch the credit table, and move up only once a workflow is predictable.
What we found
Luma AI is a Palo Alto company, Luma AI, Inc., that started out turning iPhone footage into 3D scenes and is now mainly a video generator sold as a creative agent. Its consumer video product, Dream Machine, arrived in June 2024; on 5 March 2026 Luma relaunched the service as Luma Agents, built on a new image model, Uni-1, which plans a brief and routes each step to Luma's own Ray models or to third-party ones such as Veo, Kling, Seedance, GPT Image and ElevenLabs. The 3D Capture app is still in the App Store, last updated in January 2026, but its Genie text-to-3D feature was switched off on 1 January 2026.
The current video model is Ray3.2, released on 9 June 2026. It allows up to 16 keyframes inside one clip, generates in HDR with 16-bit EXR export, runs to 20 seconds at 1080p and, for the first time, is offered through an API. Later additions include Skills for saving a workflow, Scenes, Variants for resizing and translating a finished ad, and an October 2026 connector that opens Claude Motion animations in Luma.
Plans are priced in credits that do not roll over: Plus is $30 a month for 10,000 credits, Pro $90 for 40,000 and Ultra $300 for 150,000, with yearly billing saving up to 20%. No free plan appears on the pricing page. Credit cost is where users struggle. On Trustpilot, read on 14 October 2026, Luma scored 1.5 from 68 reviews, 86% of them one-star, on a profile Luma has not claimed; the recent complaints describe failed or unwanted generations eating a month's credits.
Funding is not the worry. A $900 million Series C in November 2025 was led by Humain, the Saudi PIF-owned AI company, with AMD participating, and paired with a planned 2GW compute cluster in Saudi Arabia. Luma opened a Riyadh office in February 2026 and has announced an Arabic interface. It suits agencies and studios that want directable video in one workspace and can budget credits carefully.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
Notable facts
- Luma AI received over 1 million video generation requests in its first 10 days of public access to Dream Machine.
- The company's 3D capture technology was used by product teams at IKEA and Amazon to generate 3D models of physical products without professional photography studios.
- Luma's founding team includes PhD graduates from UC Berkeley's NeRF research group, which invented the Neural Radiance Field technique.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including Luma AI's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Luma AI unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Luma AI directly →
Spotted something out of date? Suggest an update →
Alternatives to Luma AI
More in Video & Audio
