You have successfully installed version 11.0.0 of the Gurobi Optimizer.

You’ll find instructions for setting up your Gurobi license in this document, as well as a list of supported platforms, and Release Notes and guidelines for converting existing Gurobi code to run with this new version. Once you have finished browsing this document, you should refer to the Detailed Release Notes, the Getting Started Knowledge Base article, the Example Tour, the Reference Manual, or the Remote Services Reference Manual to get started using the Gurobi Optimizer.


Obtaining Your License

To obtain a Gurobi 11 license, you will need to visit the Gurobi User Portal. If you are a commercial user under maintenance, you should see your Gurobi 11 license under Licenses. If you would like to request a free academic license, you can do so from the Licenses Request section. Once you have a license in the Gurobi User Portal, you will need to follow the instructions for installing or downloading a license by clicking on the Install or Open button next to your license.


Release Notes for Version 11.0

In addition to performance enhancements, Gurobi 11.0 adds the following new features:

  • Mixed-Integer Non-Linear Programming (MINLP) Problems: Gurobi 11 offers solving non-linear models using spatial branch-and-bound and outer approximation as an alternative to the static piecewise-linear approximation of non-linear constraints supported in earlier versions. See the discussion of function constraints with dynamic piecewise-linear approximation for more information.
  • Default behavior of NonConvex parameter changed: now, in default settings, Gurobi will solve models with non-convex quadratic constraints.
  • Models can now be copied between environments.
  • Better control over the concurrent LP optimizer with ConcurrentMethod.
  • New parameters:
  • Numerous changes to gurobipy:
    • Installation and packaging changes.
    • New classes, methods and properties.
    • Other enhancements.

Please refer to the Detailed Release Notes for more details, including supported platforms and code conversion notes. All bug fixes listed by version can be found on our website.


Supported Platforms for Version 11.0

Platform (port)Operating SystemCompilerNotes
Windows 64-bit (win64)Windows 10, 11, Windows Server 2016, 2019, 2022Visual Studio 2017-2022Use gurobi_c++md2017.lib (e.g.) for C++
Linux® x86-64 64-bit (linux64)Red Hat® Enterprise Linux 8, 9GCC >= 8.5 
SUSE® Enterprise Linux 15
Ubuntu® 20.04, 22.04
Amazon Linux 2
macOS 64-bit universal2 (macos_universal2)12 (Monterey), 13 (Ventura), 14 (Sonoma)Xcode 13/14 
Linux® arm64 64-bit (armlinux64)Red Hat® Enterprise 8, 9GCC >= 8.5 
SUSE® Enterprise Linux 15
Ubuntu® 20.04, 22.04
Amazon Linux 2
AIX® 64-bit (power64)AIX 7.2, 7.3XL C/C++ 9Note that Gurobi 11 will be the last supported version for AIX. Due to limited Python support on AIX, this port does not include the Interactive Shell or the Python libraries.


Additional Supported Platform Information

Gurobi 11.0 supports the following language/platform versions:

Python3.8, 3.9, 3.10, 3.11, 3.12
JDK8, 11, 17


Bug Fixes

All bug fixes listed by version can be found on our website.


Changed in Gurobi 11.0

Linux GNU C library (glibc) dependency

  • Linux x86_64: The Gurobi Optimizer now requires a minimum glibc version of 2.17. For version 10.0, the minimum required glibc version was 2.3.4.

    The list of supported platforms above takes these changes into account.


New features affecting all APIs

Mixed-Integer Non-Linear Programming (MINLP) Problems

The Gurobi Optimizer can now use spatial branching and outer approximation to solve models with non-linear functions, instead of using static piecewise-linear approximations. See the discussion of function constraints with dynamic piecewise-linear approximation for more information.

Default behavior of NonConvex parameter

The default behavior of the NonConvex parameter has changed. Previously, you had to explicitly set this parameter to value 2 if you wanted to solve a model with a non-convex quadratic objective or constraint. Now it will solve such models by default. To revert to the previous behavior, set the parameter to 1.

Copying models between environments

It is now possible to copy a model from one environment to another, using the following new routines or signatures:

The primary use case for copying models is when performing optimization on multiple, related versions of the same problem on multiple threads. Multi-threading within an environment is not supported.

For Compute Server users, note that you can copy a model from a client to a Compute Server environment, but it is not possible to copy models between two Compute Server environments.

Better control over the concurrent LP optimizer Option 5 of the Method parameter has been deprecated in favor of the new ConcurrentMethod parameter. The latter allows you to control which methods are run concurrently with the deterministic (Method=4) or non-deterministic (Method=3) concurrent optimizer. The deprecated setting Method=5 can now be replicated by choosing Method=4 along with ConcurrentMethod=3. To give another example, non-deterministic concurrent execution of the primal and dual simplex algorithms can be selected using Method=3 along with ConcurrentMethod=3.

Change the Threads parameter when resuming optimization If an optimization is interrupted (e.g., due to a time limit), the user can now modify the Threads parameter in order to change the number of threads to be used when the solving process is resumed with another call to optimize.



The distributed tuning can use remote workers in a dynamic way. With the parameter TuneDynamicJobs you can control the size of a dynamic set of remote workers that are used in parallel. These workers are used for a limited amount of time and afterwards potentially returned so that they are available for other remote jobs.

Starting from Gurobi 11, we include in the results the first set of parameters used to solve the model (base settings). This result is stored as first setting in the tune results. Hence, to get the best tuned set of parameters, you need to request at least 2 sets (using TuneResults) and retrieve the set with index 1.

The tuner now supports the log callback. You can add a log callback to the environment and retrieve all the log messages coming from the tuner.

Piecewise-linear approximation The default behavior of parameter FuncPieces, the main parameter for piecewise-linear approximation of function constraints, is changed. The new default for version 11.0 uses the relative error approach for the approximation, while for version 10.0 or earlier it mainly used the number of function constraints to set the total number of pieces.

New parameters The following parameters are new in Gurobi 11.0:

  • SolutionTarget to specify the solution target for LP.
  • FuncNonlinear to control whether general function constraints shall be treated as nonlinear functions or via piecewise-linear approximation.
  • MixingCut to control the generation of Mixing Cuts.
  • TuneDynamicJobs to enable distributed tuning using a dynamic set of workers.
  • TuneUseFilename to use the model file names as model names during tuning (Command-line only).


Changes to gurobipy

Installation and packaging changes

  • gurobipy and its file are no longer distributed with the Gurobi installers. gurobipy must now be installed into Python environments using pip or conda.
  • The gurobi conda packages have been updated so that pip can now recognize when gurobipy has already been installed into an environment by conda. As a result, using pip in a conda environment to install a package that depends on gurobipy (such as gurobi-machinelearning or gurobipy-pandas) no longer leads to a duplicate installation of gurobipy.
  • Type hints for gurobipy are now included in both the pip wheels and conda packages. For Gurobi version 11, users should not install the gurobipy-stubs package as it will override the type annotations packaged with gurobipy. The gurobipy-stubs package will not be maintained for future versions of gurobipy.

New classes, methods, and properties

  • SOS objects have a new index property returning the index of the SOS constraint in the model.
  • MConstr and MQConstr objects now implement the class method fromlist which has the same behavior as the corresponding method on the MVar class.
  • New methods hstackvstack, and concatenate allow matrix-friendly objects to be joined together in the same way as the corresponding numpy functions.
  • A new MGenConstr class has been added for holding an array of general constraints. For version 11, only the addGenConstrIndicator method is able to return objects of this type.
  • Env objects now have an additional method getParam which returns the current value of the given parameter in the environment.

Enhancements to existing methods

  • The Model.optimize and Model.computeIIS methods now accept any Python callable as a callback function. The TSP example code has been updated to demonstrate the use of a callable class.
  • The matrix-friendly API is now fully integrated with callback methods: Model.cbLazyModel.cbCut, and Model.cbSetSolution now accept matrix-friendly API objects.
  • When constructing a set of linear constraints using Model.addMConstr or Model.addConstr with the matrix-friendly API, user-defined constraint names can be passed as a list or ndarray.
  • Model.addGenConstrIndicator and the overloaded » syntax for indicator constraints now handle matrix-friendly linear expression objects. This allows indicator constraints to be built in a vectorized manner (including broadcasting behavior).
  • Users can now explicitly disable the default environment by setting the environment variable GUROBIPY_ALLOW_DEFAULTENV to 0 before starting a Python program. With this setting, environments must be managed explicitly using Env objects in user code. Calls to Model() without an Env object and global function calls such as setParam will raise a GurobiError if this setting is used.

Changes that may require updating your code

  • Model.getAttr and Model.setAttr now raise an exception if the modeling objects passed to them do not belong to the model. Previously, the attribute values would be returned or set for a variable on the model with the same index as the modeling object which is typically not the desired behavior.

  • A call to Model.addGenConstrIndicator using size-1 matrix-friendly modeling objects, which previously returned a GenConstr object, will now return a size-1 MGenConstr. The constraint name may also change depending on the shape of the result. To revert to the previous behaviour, call .item() on matrix-friendly modeling objects passed to this function.
  • The tupledict now follows Python 3 dictionary semantics: its keysvalues, and items methods now return iterable views instead of lists. Importantly, keys no longer returns a tuplelist. If you need this behavior, you can instead call gp.tuplelist(td.keys()).
  • Env.setParam now raises a GurobiError if the user specifies a parameter which does not exist. In previous versions, such calls would print a message but would fail without an exception.
  • Gurobi log messages now go to the `gurobipy’ logger instead of `gurobipy.gurobipy’. If you use Python’s logging library to collect Gurobi log messages, handlers should be attached either to the root logger or to logging.getLogger(“gurobipy”).
  • Calling the sum() method of an MVarMLinExpr, or MQuadExpr which contains no elements (i.e. has at least one zero-length dimension) now returns an array of zeros of the appropriate shape instead of raising an exception. This matches the behavior of numpy arrays.
  • Passing too many keys (more than the length of the contained tuples) to any of the pattern matching methods (tuplelist.selecttupledict.selecttupledict.sum, and now raises an IndexError. This change almost exclusively affects cases which returned an empty set in previous versions. The exception is where the additional indices are wildcards. i.e. for a tuplelist containing 2-tuples:“*”, 2, “*”) now raises an exception, while in previous versions it gives the same (non-empty) result as“*”, 2).
  • The gurobi Python package alias has been removed from the conda packages on all platforms. Any usage of import gurobi must be replaced with import gurobipy.

Deprecated functionality

If you are upgrading from a previous version of Gurobi, we recommend first running your code with Gurobi 10 and warnings enabled to catch deprecations in gurobipy. Fixing these deprecated usages will help to keep compatibility for Gurobi 11 and later versions. Warnings can be enabled by running your code with the -X dev or -W default flags. See the Python Development Mode or warnings package documentation for further details.

In Gurobi 11, the following usage is deprecated and will be removed in a future version:

  • Passing a tupledict to the quicksum function is deprecated as the result is inconsistent with passing a native Python dictionary to quicksum. The sum method of the tupledict should be used instead to compute the sum over the values of the tupledict.
  • The iterkeys()itervalues(), and iteritems() methods for tupledicts are deprecated. These methods are legacy syntax from Python 2. The corresponding Python 3 methods keys()values(), and items() now return iterable views and are therefore preferred.
  • The gurobipy.system() function is deprecated. os.system() from the Python standard library should be used instead.
  • The gurobipy.models() function is deprecated and should no longer be used.


Changes to the Java package


We’ve made essential changes in the package structure to allow publishing to Maven Central. Please take note of the following updates:

  • Package Name Change: The Java package name has been updated from gurobi to com.gurobi.gurobi. If you are integrating our Java package, ensure to update your import statements accordingly.


Changes to the MATLAB interface

On the macOS platform, the Gurobi mex interface is now also available for native Apple silicon MATLAB (mexmaca64).


Compute Server, Cluster Manager, and Instant Cloud

Detailed release notes for Compute Server, Cluster Manager, and Instant Cloud can be found here


License Agreement

Note that this software is covered by the Gurobi End User License Agreement. By completing the Gurobi installation process and using the software, you are accepting the terms of this agreement.

Thank you for using Gurobi products!

Try Gurobi for Free

Choose the evaluation license that fits you best, and start working with our Expert Team for technical guidance and support.

Evaluation License
Get a free, full-featured license of the Gurobi Optimizer to experience the performance, support, benchmarking and tuning services we provide as part of our product offering.
Academic License
Gurobi supports the teaching and use of optimization within academic institutions. We offer free, full-featured copies of Gurobi for use in class, and for research.
Cloud Trial

Request free trial hours, so you can see how quickly and easily a model can be solved on the cloud.