NONMEM® versions up through VI are the property of the Regents of the University of California, but ICON Development Solutions has exclusive rights to license their use. NONMEM® 7 is the current version of the software and is the property of ICON Development Solutions. 

The software consists of three parts.

• The NONMEM® program itself is a very general (non-interactive) regression program that can be used to fit models to many different types of data. With version 7, Monte Carlo expectation-maximization and Markov Chain Monte Carlo Bayesian methods have been added to the classical likelihood methods available in previous versions. NONMEM® can be used to simulate data as well as fit data.
• PREDPP is a powerful package of subroutines that can be used by NONMEM® to compute predictions for population pharmacokinetic and pharmacodynamic data. Use of PREDPP obviates the need for the user to code kinetic-type equations and it also allows complicated patient-type data to be easily used. However, the user can also directly and very generally code prediction-type, likelihood or -2 log-likelihood equations, and thus a great variety of different types of models can be used.
• NM-TRAN is (non-interactive) preprocessor, allowing control and other needed inputs to NONMEM/PREDPP to be specified in a user friendly manner, with quick and comprehensive detection of a variety of errors that the user may have made in so doing.

NONMEM® 7 is programmed in Fortran 90/95 code and can be used with hardware and Window, Linux or Solaris operating system incorporating a Fortran 90/95-compliant compiler. Since some NONMEM® runs can take considerable time, perhaps many hours or even days, depending on the speed of the computer and the size of the problem, it is advisable to use a fast machine with at least 1.0 Gb of available memory. For multiprocessor and multi-core environments 3-4 Gb of memory may be needed to accommodate several simultaneous NONMEM® runs.
 

• Improved incidence of success in problems using the first-order conditional estimation method.
• Improved incidence of completion when using the “Super Problem” feature.
• Exact likelihood maximisation methods, such as importance sampling expectation maximisation (EM), and stochastic approximation EM.
• Full three stage hierarchical Markov Chain Monte Carlo (MCMC) methods.
• Additional result files, with number of significant digits selectable by the user, and which can be easily read by post-processing programs.
• Numbers of data items per data record increased to 50.
• Label names may be as large as 10 characters.
• Initial parameter entries in the control stream file may be of any numerical format.

The NONMEM program is available on CD ROM, which together with the documentation and all updates and additions to the program and documentation, will be delivered for a license royalty fee to be paid annually. This fee is subject to change from year to year, and at each anniversary the licensee at its option may choose not to renew the license.