Overview
Luma AI, Inc. is a private AI research company building multimodal generative models for image, video, and audio creation. Its current models are Ray3.2 for video (released June 2026) and Uni-1 for image and multimodal reasoning. The Luma App at app.lumalabs.ai is the consumer workspace, with agentic "Luma Agents" workflows across video, image, audio, and text, including third-party model orchestration.
Video: text/image-to-video, Modify Video V2 (transform footage up to 20s at 1080p), multi-keyframe direction (up to 16 keyframes), reframe for any aspect ratio, Draft Mode, native HDR + EXR export. Image: text-to-image, reference-guided generation (up to 9 references), natural-language editing.
Credit-based pricing, verified live on the official pricing page 2026-10-04: Plus $30/month (10,000 credits), Pro $90/month (40,000 credits, 4x Luma Agents usage), Ultra $300/month (150,000 credits, 15x Luma Agents usage); Team and Enterprise via contact sales; yearly billing saves up to 20%. Commercial use is included on all paid plans. The official pages note free trial credits on all plans; no standalone permanent free tier is stated.
Key Features
- Text/image-to-video (Ray3.2)
- Modify Video V2
- Multi-keyframe direction
- Reframe for any aspect ratio
- Draft Mode
- Native HDR + EXR export
- Text-to-image and natural-language image editing (Uni-1)
- Luma Agents workflows
- Developer API (Python + JS/TS SDKs)
- Audio tools (TTS, SFX, music, transcription)
- Team features (projects, SSO, shared credits)
