Lexsi Labs, an Aurionpro company (formerly AryaXAI Alignment Labs), has announced two major advancements aimed at transforming how enterprises utilize AI for structured/tabular data.
Launch of Orion-MSP: A new state-of-the-art (SOTA) tabular foundation model designed to deliver enterprise-grade accuracy and scalable deployment across any predictive task involving structured data.
Release of TabTune Open-Source Library: The first open-source tool that standardizes fine-tuning and inferencing for leading tabular foundation models, including Orion-MSP.
These developments represent a key milestone in Lexsi Labs’ mission to build “Safe SuperIntelligence” systems that are transparent, aligned, and trustworthy for enterprise use.
Targeting industries such as banking, insurance, and healthcare, Orion-MSP allows organizations to predict tasks in zero-shot mode using a single API-based model, eliminating complex multi-pipeline workflows. Enterprises can also fine-tune models effortlessly to achieve optimal performance within just a few lines of code.
Vinay Kumar, Founder & CEO of Arya.ai and Lexsi Labs, stated,
“While LLMs, LRMs, and LVMs have advanced zero-shot capabilities in their domains, Orion-MSP brings this power to tabular data. It delivers top mean-rank accuracy across multiple benchmarks, outperforming existing tabular foundation models and classical machine learning approaches.”
He added that TabTune serves as an ideal companion tool for enterprises and practitioners to infer or fine-tune tabular foundation models, with built-in metrics for fairness and conformity. Both Orion-MSP and TabTune are being released under the MIT license, ensuring open, unrestricted use for the global AI community.