Dear Dynare friends,

We are pleased to announce the release of Dynare 6.4. This maintenance release fixes various bugs.

The Windows, macOS, MATLAB online and source packages are available for download at the Dynare website. This release is compatible with MATLAB versions ranging from 9.5 (R2018b) to 25.1 (R2025a), and with GNU Octave versions ranging from 7.1.0 to 10.2.0 (NB: the Windows package requires version 10.2.0 specifically).

Here is a list of the problems identified in version 6.3 and that have been fixed in version 6.4: * Several issues with the perfect_foresight_with_expectation_errors command: when used with no endval/endval(learnt_in=1) block but an endval(learnt_in=t) block with t>1, convergence could fail if homotopy was neededcombined with homotopy_marginal_linearization_fallback, it would return incorrect resultsit would crash with homotopy_marginal_linearization_fallback option if there were several distinct shocks or endval blocks with the learnt_in option * Using the model_diagnostics command with the bytecode option of the model block or the model_options command would crash if the static Jacobian was singular * The discretionary_policy command with the noprint option would crash when encountering a problem instead of continuing * Using the ramsey_model command with models declared as linear would crash during preprocessing when there is a lead/lag greater than 1 on endogenous variables or any lead/lag on exogenous variables * Using the identification command with the slice sampler would crash if the slice_initialize_with_mode option was used * The particle filter option resampling_method of the estimation command was broken * The output=third preprocessor option would not work if no computational commands were present in the .mod file

As a reminder, the list of new features introduced in versions 6.x can be found in the release notes for 6.0.

On behalf of the Dynare Team,