Pumpkin is a new combinatorial optimisation solver that is under active development at the ConSol Lab at TU Delft.
The contributors (all at TU Delft) of Pumpkin are (in alphabetical order): Emir Demirović, Maarten Flippo, Imko Marijnissen, Konstantin Sidorov, and Jeff Smits.

Pumpkin has a FlatZinc frontend, but can also be used to solve SAT and MaxSAT problems. Support for CPMpy and XCSP is planned for the near future.
The solver is written in pure Rust and also exposes a Rust API which allows users to express models with predefined constraints and implement custom branching strategies.

The solver is based on the (lazy clause generation) constraint programming paradigm. 
It supports integer variables, the FlatZinc built-in constraints for integer and boolean variables, and the cumulative global constraint. 

A unique feature of Pumpkin is that it can produce a certificate of unsatisfiability as part of the solving process without significant overhead. See our CP’24 [1] paper for more details. 

The solver is reasonably competitive, although depending on the problem, other state-of-the-art solvers may be better. We are planning many improvements for the near future.

For more information, the repository can be found at https://github.com/ConSol-Lab/pumpkin. We welcome questions in the Discussion section of GitHub.

[1]. A Multi-Stage Proof Logging Framework to Certify the Correctness of CP Solvers. Flippo, Sidorov, Marijnissen, Smits, Demirović. CP'24.