Stan

Health v2 · maintenance only

Adoption not part of the score

2761 stars · 387 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Stan is a C++ probabilistic programming library for full Bayesian inference via NUTS/Hamiltonian Monte Carlo, approximate inference via ADVI, and penalized maximum likelihood via L-BFGS. It is built on the Stan Math automatic differentiation library and offers interfaces from R, Python, Julia, and the command line.

Use cases

• fit bayesian hierarchical models to data
• run MCMC sampling on a custom statistical model
• do variational inference for approximate posteriors
• estimate model parameters with maximum likelihood
• write probabilistic models in a declarative modeling language
• fit time-series or multilevel regression models

When to choose

• you need full Bayesian posterior inference with state-of-the-art HMC sampling
• you want a mature, well-documented probabilistic programming language with multi-language interfaces
• you need automatic differentiation for custom statistical models
• you are doing academic or production statistical modeling in R, Python, or Julia

When to avoid

• you only need simple frequentist regression available in base statistical packages
• you cannot install a C++17 toolchain
• you need deep learning or GPU-heavy neural network training rather than statistical inference
• you want a pure-Python or pure-R solution without compiled dependencies

Facets

library · maturity stable

machine-learning data-science math simulation machine-learning data-science cross-platform cpp python jvm cli bayesian-inference probabilistic-programming mcmc hamiltonian-monte-carlo variational-inference automatic-differentiation statistical-modeling statistics

Member repositories

For agents

markdown · JSON · MCP: product_card(name="stan-dev/stan")