Pumas 2.9 Adds Faster Parallel Drug-Model Estimation
The update extends the Pumas 2.8 release with multithreading and distributed computing, though the source gives no benchmark figures.
Updated on Nov. 6, 2026 in Software

PumasAI released Pumas 2.9, a scientific software package for clinical pharmacology and pharmacometrics, building on its prior 2.8 version. The release is a shipped software update, not an announced or research-stage capability.
Why it matters
Pharmacometrics software, which models how drugs behave in the body, is increasingly built around parallel computing and AI-assisted analysis. Pumas 2.9 extends that trend by focusing on speed, reproducibility, and stability for complex estimation problems.
The release notes list enhancements to the experimental Monte Carlo Expectation Maximisation and Variational Expectation Maximisation estimation methods, plus new multithreading and distributed computing with load balancing. The source gives no benchmark figures, so the scale of improvement is unquantified.
The players
PumasAI
An AI-powered clinical pharmacology and pharmacometrics company that develops the Pumas software for modeling drug behavior.
The details
Pumas 2.9 targets reproducible inference when calculations run across parallel computing environments, such as multiple processor cores or several machines. Multithreading lets one program run several tasks at once on a single machine, while distributed computing spreads work across networked machines; intelligent load balancing assigns work so no single machine sits idle. The update also adds support for ordered parameter domains, which constrain model parameters to a defined sequence, and expands covariate selection, error modeling, and diagnostics used in drug-response models.
Timeline
Pumas 2.9 was released on November 6, 2026.
The Tech Race
Pumas 2.9 continues a version line that already added estimation and simulation features in 2.8, placing PumasAI's pharmacometrics tooling on an incremental roadmap of parallelism and reproducibility. The source does not name competing pharmacometrics packages, so its positioning relative to rivals cannot be assessed from this release alone.
Researchers and pharmaceutical modelers who use Pumas can now install version 2.9, which adds multithreading, distributed computing, and expanded diagnostics. The source does not state pricing, licensing changes, or a timeline for wider availability beyond the release itself.
The takeaway
Pumas 2.9 signals that pharmacometrics software is moving toward faster parallel computation without requiring users to rewrite their workflows. Watch for published benchmarks from PumasAI quantifying the speed and stability gains, which the release notes currently leave unstated.
Further reading
For broader context on how scientific computing tools are evolving, see the Software section.


