Mplus Statistical Modeling Software (Muthén & Muthén): Features and Use

Mplus is a statistical modeling program from Muthén & Muthén that fits a wide range of models, estimators and algorithms, and is used by researchers for latent variable work such as structural equation, multilevel and mixture models. This page summarizes what the vendor says it covers, where it runs, how it is obtained, and which alternatives to weigh. It is not the game feature “Mythic+” (Mplus in World of Warcraft searches), and it is not a tutorial. Facts below were checked against the vendor’s pages on 2026-11-07.

What Mplus is

Per the vendor’s features page, Mplus is a statistical modeling program with a wide choice of models, estimators and algorithms, plus graphical displays of data and results. It handles cross-sectional and longitudinal data, single-level and multilevel data, data with observed and unobserved heterogeneity, data with missing values, and Monte Carlo simulation. The latest release listed on the version-history page is Mplus Version 9.1.1 (16 September 2026), which fixes minor problems found since Version 9.1 (May 2026).

Who runs it and what it covers

Mplus is developed and sold by Muthén & Muthén at statmodel.com. The table maps model families to what the vendor’s features page lists; it is a summary of the vendor’s own description, not an independent test.

Area	What the vendor says Mplus offers
Outcome types	Continuous, censored, binary, ordered categorical, unordered (nominal) categorical, count, or combinations
Continuous latent variables	Factors, random effects, frailties, genetic liabilities, latent responses for missing data
Categorical latent variables	Latent classes, latent trajectory classes, mixture components
Regression	Linear, tobit, probit and logistic, multinomial logistic, Poisson and negative binomial (with or without zero inflation), including within- and between-cluster variation in multilevel data
Data structure	Cross-sectional, longitudinal, single-level, multilevel, missing values, Monte Carlo simulation

For the methods themselves, see the CASRAI guides on path analysis, confirmatory factor analysis and full information maximum likelihood (FIML), which handles missing data without a separate imputation step.

How to get and run it

Where to find it: the vendor’s online order page and system requirements page are on statmodel.com, reached from the site’s menu. What the vendor’s pages support:

1. Check the system requirements (table below).
2. Try the demo version, which the vendor describes as identical to the regular version except that the number of variables is limited. The vendor recommends testing it before buying.
3. Buy through the vendor’s order page. Muthén & Muthén accepts orders only from end users; payment is by credit card, PayPal, or an advance check on a U.S. bank; wire transfers are not accepted; U.S. purchase orders over $250.00 are accepted and international purchase orders are not; all sales are final. A student license addendum is published on the order page.
4. Write an input file. The version history names the input commands DATA, VARIABLE, ANALYSIS, MODEL, MODEL CONSTRAINT, MODEL TEST, MODEL PRIORS, OUTPUT, SAVEDATA, DEFINE, PLOT, MONTECARLO, DATA IMPUTATION and DATA MISSING.
5. Run the file and read the output it produces. Command syntax is documented in the vendor’s Mplus User’s Guide.

Platform requirements

Item	Vendor statement
Memory and disk	At least 1 GB RAM and 120 MB disk; memory needs vary with the analysis and data size
Windows	Windows 10 and 11 (Intel apps run under emulation on Windows 11 Arm PCs)
macOS	10.11 (El Capitan) through 15 (Sequoia); does not run natively on Apple M-family chips, and runs through Rosetta with reportedly reduced performance
Linux	64-bit, tested on Ubuntu, RedHat, Fedora, Debian and Gentoo; command line only, with no special cluster capabilities

Strengths and limits

• Breadth in one program: the features page lists categorical and continuous latent variables, many outcome types and multilevel data together.
• Try before buying: a demo with a variable limit exists.
• Purchase terms are strict: end users only, sales final, no wire transfers, no international purchase orders.
• Apple silicon: not native, per the vendor.
• Linux: command line only.

Alternatives

Option	What its own page says
lavaan (R package)	Described as a free open-source package for latent variable modeling in R; examples cover path models, confirmatory factor analysis, full structural equation models and growth curve models (lavaan.ugent.be, checked 2026-11-07)
IBM SPSS Amos	Structural equation modeling software that extends regression, factor and variance analysis; supports path analysis, confirmatory factor analysis, Bayesian estimation, categorical and censored data, and latent class analysis; Windows only (ibm.com, checked 2026-11-07)

For general-purpose packages, see SPSS vs R and SPSS vs Stata.

FAQ

Does Mplus run on a Mac?

Yes on Intel Macs running macOS 10.11 through 15. The vendor states it does not run natively on Apple M-family chips; it runs via Rosetta, with performance reportedly decreasing.

Is there a free version of Mplus?

There is a free demo that is identical to the regular version except that the number of variables is limited. The full program is sold by Muthén & Muthén.

Is Mplus the same as Mythic+ in World of Warcraft?

No. This page covers the statistical program only.

Sources

• Muthén & Muthén, Mplus features: https://www.statmodel.com/features.shtml (checked 2026-11-07)
• Muthén & Muthén, order online: https://www.statmodel.com/orderonline/ (checked 2026-11-07)
• Muthén & Muthén, system requirements: https://www.statmodel.com/sysreq.shtml (checked 2026-11-07)
• Muthén & Muthén, version history: https://www.statmodel.com/verhistory.shtml (checked 2026-11-07)
• lavaan project site: https://lavaan.ugent.be/ (checked 2026-11-07)
• IBM SPSS Amos: https://www.ibm.com/products/structural-equation-modeling-sem (checked 2026-11-07)