Base44 Becomes First App-Creation Platform to Launch Its Own Proprietary LLM “Base 1”, Marking a Major Milestone in the Company’s Technology Vision

Ownership of the model gives Base44 direct control over compute and inference spend, expected to result in a structurally stronger margin profile over time

Base44, the AI-powered platform that enables anyone to create production-ready applications using natural language, announced the launch of its first proprietary AI model, now in production and actively serving users on the platform. The milestone marks a defining moment in Base44’s evolution, from an AI-powered platform to a full-stack AI company with deep, end-to-end capabilities spanning data, infrastructure, training, and deployment.

The custom LLM was developed and optimized using Base44’s vast and unique dataset generated from tens of millions of real user interactions on the platform. The result is a model already demonstrating competitive performance against the foundation models it replaces, with an opportunity to surpass them as training continues on Base44’s unique and growing dataset.

“As the first app-creation platform to launch its own proprietary AI model, this is a big step for us as a company,” said Maor Shlomo, Founder and CEO of Base44. “Having our own model means we can continue to improve it over time, making us less dependent on external vendors. This is not a trivial effort. It requires deep expertise, significant infrastructure, and access to the kind of rich, diverse data that most organizations simply do not have.”

Also Read: AiThority Interview with Matej Bukovinski, Chief Technology Officer at Nutrient

The new model represents a very ambitious custom model development effort underway in the industry. Unlike most companies that train models for narrow, single-purpose use cases, Base44 has built a general-purpose AI agent capable of handling both conversation and coding tasks, including complex multi-turn interactions, tool usage, runtime management, and backend operations.

Base44’s ability to build and deploy its own model is the result of deliberate investment across several key areas:

• Talent: Wix’s world-class data science and research team has developed deep expertise through the process of training Wix Harmony and is applying that expertise to development of the Base44 LLM .
• Data: Base44 has accumulated a tremendous amount of high-quality data from production traffic and user interactions, giving the model unique and powerful training signals specific to Base44 use cases.
• Infrastructure: The company has built robust sandboxing capabilities, evaluation frameworks, and experimentation pipelines that allow different agent configurations to be tested in real time and deployed on live traffic.
• Built by: Wix and Base44 in collaboration with top inference providers to ensure it runs with the performance, reliability, and scale that users expect.

Owning the model stack gives Base44 direct control over one of the largest cost drivers in AI-native businesses, compute and inference spend, while continuing to improve the quality of outcomes for users. This milestone is a meaningful step for Base44 and is expected to structurally improve profitability over time as the model is gradually implemented.

The team sees this moment as the first of many milestones, with larger models, faster iteration cycles, and deeper product integration on the horizon.