Pumas 2.8 Sets a New Standard for Clinical Pharmacology and Pharmacometrics Technology with Embedded AI and Next Generation Estimation

DOVER, Del. — DOVER, Del. (February 17, 2026) – PumasAI, a science-first organization that turns data into life-saving decisions faster, today announced the release of Pumas version 2.8 to bring significant advancements in clinical pharmacology and pharmacometrics. The latest version integrates the recently launched PumasAide, enabling artificial intelligence (AI) coding agents to connect directly to live Pumas sessions, alongside powerful new estimation methods and expanded modeling capabilities.Within Pumas 2.8, pharmaceutical and biotech teams can perform nonlinear mixed effects (NLME) modeling, optimal design, Bayesian analyses, Expectation Maximization based analysis, clinical trial simulations, machine learning, non-compartmental analysis, and bioequivalence studies within one integrated software environment. By consolidating workflows inside a single platform, organizations can avoid productivity loss, reduce the cost of maintaining multiple software systems, and access the latest innovative technologies in one place. The result is greater efficiency and more time focused on scientific thinking. “This release brings true AI integration to pharmacometrics,” said Dr. Vijay Ivaturi, Co-founder and CEO of PumasAI. “By connecting intelligent coding agents to real-time modeling sessions, we are transforming how scientists explore, build, and communicate their results.”Major features in the Pumas 2.8 release include:Integrated PumasAide: Connect coding agents directly to Pumas sessions to generate, execute, and refine models within the user’s coding environment.Monte Carlo Expectation Maximization (MCEM): A new estimation method available for regular models defined using @model, expanding flexibility for complex NLME analyses.Variational Expectation Maximization (VEM): An additional estimation method for regular @model workflows, providing alternative approaches for efficient parameter estimation.FiniteMarkov Distribution: Enables modeling of discrete time finite state Markov processes directly within Pumas, supporting advanced disease progression and state transition analyses.Platform Enhancements: Additional improvements and refinements across modeling, simulation, and performance, as detailed in the official Pumas 2.8 release notes.Dr. Ivaturi added, “Clinical pharmacology and pharmacometrics are entering a new era where intelligent systems work alongside scientists to amplify insight, not replace it. With Pumas 2.8, we are building the foundation for a future where modeling, simulation, and AI operate as one continuous engine for discovery, helping teams move faster, think deeper, and ultimately bring better therapies to patients sooner.”Pumas 2.8 is available on fully validated (SOC-2, CFR Part 11) JuliaHub cloud environment or for desktop use. Students, professors and researchers can gain complimentary access to Pumas and other powerful modeling tools for non-commercial research and education by emailing their interest to [email protected].For more information or to schedule a demonstration, visit: https://pumas.ai/our-products/pumas.