Ten-Dimensional Neural Network Emulator for the Nonlinear Matter Power Spectrum

We present gokunemu, a ten-dimensional neural network emulator for the nonlinear matter power spectrum, designed to support next-generation cosmological analyses. Built on the Goku N-body simulation suite and the t2n-muse emulation framework, gokunemu predicts the matter power spectrum with ∼0.5% average accuracy for redshifts 0≤z≤3 and scales 0.006≤k/(h Mpc^{-1})≤10. The emulator models a 10D parameter space that extends beyond Λ-cold dark matter (ΛCDM) to include dynamical dark energy (characterized by w_{0} and w_{a}), massive neutrinos (∑m_{ν}), the effective number of neutrinos (N_{eff}), and running of the primordial spectral index (α_{s}). Its broad parameter coverage, particularly for the extensions, makes it the only matter power spectrum emulator encompassing the range of dynamical dark energy models preferred by recent DESI constraints. In addition, it requires only ∼2 milliseconds to predict a single cosmology on a laptop, orders of magnitude faster than existing emulators. These features make gokunemu a uniquely powerful tool for interpreting observational data from upcoming surveys such as LSST, Euclid, the Roman Space Telescope, and CSST.