PumasAI has released Pumas 2.9, the latest version of its clinical pharmacology and pharmacometrics software. The company said the update improves computational performance, estimation stability and reproducibility, and supports more sophisticated pharmacometric workflows. The release was announced on 6 November 2026 from Dover, Delaware.
PumasAI describes itself as an AI-powered clinical pharmacology and pharmacometrics company. It develops the Pumas suite, an integrated modelling and simulation platform used across the drug development lifecycle. Its scientists also provide consulting to pharmaceutical companies in clinical pharmacology, model-informed drug development (MIDD) and pharmacometrics.
The new version builds on the artificial intelligence capabilities that the company introduced in Pumas 2.8 through PumasAide. According to PumasAI, version 2.9 strengthens the scientific engine behind model-informed drug development. The changes cover estimation methods, model flexibility, diagnostics, distributed computing and user experience.
“Pumas was built to help scientists solve increasingly complex problems without adding unnecessary complexity to their workflows,” said Dr Vijay Ivaturi, co-founder and chief executive of PumasAI. “Pumas 2.9 continues that mission by strengthening the core modeling platform with faster computation, greater stability, and expanded capabilities that allow researchers to spend more time generating scientific insight.”
Estimation and modelling changes in the pharmacometrics software
The main changes concern two experimental estimation methods, Monte Carlo expectation maximisation (MCEM) and variational expectation maximisation (VEM). The company said the latest work on both methods improves stability and reliability. It also said the methods now offer better support for increasingly complex nonlinear mixed-effects models.
Modelling flexibility has also been extended. Pumas 2.9 adds support for ordered parameter domains and enhanced covariate selection. The release also includes richer prediction outputs, expanded error model capabilities and additional inference tools for building pharmacometric models.
Performance and reproducibility
On the computing side, the release adds new multithreading capabilities, distributed computing enhancements and intelligent load balancing. PumasAI said these speed up simulations, diagnostics, inference and estimation in demanding modelling workflows.
The company has also made changes to deterministic calculations and to reproducible inference across parallel computing environments. It said these are intended to give more consistent scientific results. Other changes include improved diagnostics, clearer documentation, expanded error reporting and a number of workflow refinements. The company said these should help users find and resolve issues more efficiently.
A single platform for model-informed drug development
Pumas combines a wide range of analyses in one software environment. These include:
- nonlinear mixed-effects (NLME) modelling
- optimal design and Bayesian analyses
- SAEM, MCEM and VEM estimation
- clinical trial simulation and machine learning
- noncompartmental analysis and bioequivalence studies
PumasAI said that bringing these analyses onto one platform removes the need to move work between separate software packages. According to the company, this can reduce operational complexity, speed up analyses and leave more time for scientific decision-making during drug development.
“Every release of Pumas is driven by the same objective,” Ivaturi said. “Our responsibility is to give scientists the most advanced computational tools available while making those tools faster, more reliable, and easier to use. As clinical pharmacology continues to evolve, we remain committed to delivering innovations that help researchers ask better questions, explore more possibilities, and accelerate the development of new therapies.”
Availability and academic access
Pumas 2.9 is available for desktop deployment and through the JuliaHub cloud environment, which the company describes as fully validated. Students, professors and academic researchers can receive complimentary access for non-commercial education and research.
More details on the new version of the pharmacometrics software are available on the Pumas product page.
Frequently asked questions
What is Pumas 2.9?
Pumas 2.9 is the latest version of PumasAI’s clinical pharmacology and pharmacometrics platform. It brings changes to estimation methods, computational performance, reproducibility, modelling flexibility and user experience.
What estimation methods are improved in Pumas 2.9?
The release continues work on the experimental Monte Carlo expectation maximisation (MCEM) and variational expectation maximisation (VEM) methods. PumasAI said these changes improve stability, reliability and support for complex nonlinear mixed-effects models.
How does Pumas 2.9 improve computational performance?
It adds new multithreading capabilities, distributed computing enhancements and intelligent load balancing. The company said these speed up simulations, diagnostics, inference and estimation.
How can Pumas 2.9 be accessed?
Pumas 2.9 is available through the JuliaHub cloud environment and for desktop deployment.
Is Pumas 2.9 free for academic users?
Students, professors and academic researchers may receive complimentary access for non-commercial education and research.
