For a detailed changelog, see the CHANGES file in the SVN repository.

• 2010-09-24: mpmath 0.16 released. See the blog post and changelog for details.
• 2010-06-06: mpmath 0.15 released. See the blog post and changelog for details.
• 2010-02-05: mpmath 0.14 released. See the blog post and changelog for details.
• 2009-08-13: mpmath 0.13 released. See the announcement and changelog for details.
• 2009-06-09: mpmath 0.12 released. See the changelog for details.
• 2009-02-05: hcpy, an RPN calculator that uses mpmath, is available
• 2009-01-26: mpmath 0.11 released. See this blog update and the changelog for details.
• 2008-10-15: mpmath 0.10 released. See the changelog for details.
• 2008-08-23: mpmath 0.9 released. See the changelog and release announcement for details.
• 2008-08-22: The svn version of mpmath can optionally use gmpy for faster high-precision computations. See the new benchmark results.
• 2008-04-20: mpmath is 0.8 is released. See the changelog and release announcement for details.
• March 12, 2008. Released 0.7
• January 13, 2008. Released version 0.6. A new class class (mpi) for interval arithmetic has been added. Rounding for powers has been improved, and infinities/nan are supported in more places. Various small bugs have been fixed, the mpf class implementation has been simplified, and tweaks to the integer bitcount code (among other things) resulted in arithmetic now being about 20-30% faster.
• December 8, 2007. There is now a mailing list: http://groups.google.com/group/mpmath. mpmath is also being added to debian (thanks to Sandro Tosi and the Debian Python Modules Team)
• November 24, 2007. Released version 0.5. The most significant new feature is the quadts function for arbitrary-precision numerical integration using the tanh-sinh algorithm. The modulo operator and floor/ceil functions have also been implemented, and there is partial support for the special numbers inf and nan. Other small changes include a slight speedup of multiplication and division.
• November 3, 2007. Released version 0.4. String conversion has been rewritten to use integer arithmetic instead of Decimals, which makes it an order of magnitude faster; basic arithmetic speed has also been improved by 15%. Division now uses a rigorous algorithm for directed rounding (it previously got the last bit wrong with a 0.01% probability). Some functions have been added or renamed, and the unit tests have been made more exhaustive.
• October 5, 2007. Released version 0.3. All the 12 standard trigonometric, hyperbolic, inverse trigonometric and inverse hyperbolic functions are now available in mpmath and work for both real and complex numbers with arbitrary precision. A few useful nonelementary functions like gamma, erf and zeta have been implemented as well. In addition, numerous bugs have been fixed, more tests have been added, and parts of the code have been cleaned up.
• October 2, 2007. Released version 0.2. This version fixes several bugs and adds improved support for complex numbers (complex logarithms and powers are now supported).
• September 27, 2007. Released version 0.1.
• September 25, 2007. Started the Google Code project.

mpmath is based on code originally developed for SymPy in mid-2007. That code, in turn, was based on pieces of Python code written between 2005 and 2007 (the oldest being a script for computing digits of pi and e, written in early 2005).