Luma adds Uni-1.1 REST API with per-image pricing comparable to OpenAI for development workflows

Luma on May 13, 2026 opened a REST API for Uni-1.1, enabling developers to generate and edit images from text prompts and image inputs and to integrate the model directly into applications and pipelines. This release moves Uni-1.1 beyond the company’s web interface by exposing the model’s text-and-image capabilities to code-driven workflows, making it possible to embed image generation and editing into automated production stacks.

The API is available in two performance tiers: uni-1 at $0.0404 per image and uni-1-max at $0.10 per image for higher — quality outputs. Both tiers default to 2,048 — pixel output size. Requests may include up to nine reference images; Luma charges $0.003 per reference image. Luma’s API documentation supplies full details on parameters, streaming options, prompt formats and editing operations.

Beyond generation and editing, Luma says the API supports web search and built‑in reasoning as part of its text-and-image understanding. The company contrasts that approach with its web product’s agent — driven processes: the web interface runs agent workflows behind the scenes, while the API exposes underlying model primitives so developers can programmatically compose their own reasoning and iterative loops.

On public evaluation leaderboards, Uni-1.1 ranks seventh to eighth across both Arena categories, trailing competitors from OpenAI, Google and Grok Imagine. The published per-image prices and reference — image fees put Luma’s costs roughly in line with comparable offerings from Google and OpenAI, a practical consideration for teams weighing per-image economics and integration trade — offs when switching or adding providers.

For builders, the combination of 2,048 — pixel outputs, multi — image referencing and integrated search/reasoning targets iterative design and production workflows: agent — driven UIs may simplify bulk variation tasks, but the API lets teams implement equivalent iterative logic inside their own stacks. Luma says it is working on broader availability through platforms such as AWS but has not provided a launch date; immediate access is available via Luma’s REST endpoints and documentation.