The MaxSAT Evaluation 2026: Second Call for Contributions

[Jeremias Berg]

In short: Please consider submitting MaxSAT benchmarks or solvers by the 30.4.2026

See https://maxsat-evaluations.github.io/2026/submission.html for instructions on the submission procedure. Due to changes in the submission procedure following the move away from StarExec, please review the instructions, even if you have submitted to earlier iterations of the MSE. The most significant changes are also listed at the end of this e-mail. 

The 2026 MaxSAT Evaluation (MSE 2026) is the 20th edition of the MaxSAT evaluations, the primary event focused on evaluating MaxSAT solvers, organized (almost) yearly since 2006. The results will be presented at the SAT 2026 conference. The evaluation is affiliated with the FloC Olympic Games. 

The main goals of the evaluation are

• to assess the state of the art in the field of MaxSAT solvers,
• to collect and re-distribute a heterogeneous MaxSAT benchmark set for further scientific evaluations, and
• to promote MaxSAT as a viable option for solving instances of a wide range of NP-hard optimization problems.

The evaluation welcomes contributions of two types from the community at large:

• new benchmarks encoding instances of interesting NP-hard optimization problems (https://maxsat-evaluations.github.io/2026/call-for-benchmarks.html), and
• implementations of MaxSAT solvers that will be evaluated within MSE 2026 on a heterogeneous collection of benchmarks.(https://maxsat-evaluations.github.io/2026/call-for-solvers.html) 

Please consider submitting your benchmarks by 30.4.2026.

The organizers can be reached via email: maxsatev...@gmail.com.

More information can also be found on the evaluation webpage (https://maxsat-evaluations.github.io/2026/index.html)

On behalf of the organizers. 

Jeremias Berg
University of Helsinki 

Most notable changes to the MSE compared to previous years

• Solver submissions should be made by e-mail to maxsatev...@gmail.com. 
• No need to submit duplicates of the source code for every track a solver participates in; instead, participation in the different tracks is handled with runscripts. 
• Solver authors are required to provide explicit provenance information for their submitted system in the descriptions, to the extent practical.
• All solvers participating in an anytime track are evaluated under every timeout,  including an a priori unknown timeout between 900s and 1800s.