We build reduced excited-state models of non-fullerene acceptors (NFAs) for organic solar cells. All parameters come from first-principles calculations on single molecules and molecular pairs. We use range-separated hybrid DFT with a screened, optimally tuned functional (SRSH, tuned ω and α, polarizable continuum). We compute ground and excited states (TD-DFT), reorganization energies from four-point schemes, vibronic Huang–Rhys factors, dimer electronic couplings and charge-transfer state energies. The results are compared with our own transient-absorption, electroabsorption and steady-state spectra. The project has two papers. Paper 1 (NFA Y6) is a method paper. It presents a carrier- and fragment-resolved Frenkel–charge-transfer Hamiltonian built from monomer and pairwise-dimer calculations. The model reproduces the low-lying excited-state and charge-transfer level structure. It also gives two results that a whole-molecule description cannot: a decomposition of the intermolecular electron and hole couplings into site-pair (fragment) contributions, and the excited-state charge redistribution (change in dipole Δμ and polarizability Δp), which we compare with electroabsorption measurements. Paper 2 applies the same protocol to a new, unpublished acceptor (A6) and to its partners: the acceptor PC61BM and the donor P3HT (4- and 8-unit oligomers). We build two contact geometries for each interface and compute all six Marcus–Jortner parameters: electron and hole couplings, driving forces and reorganization energies. With these we test whether the measured difference between electron- and hole-transfer rates at the two interfaces follows from these parameters. So far all calculations have run on the LiU local resource Sigma (LiU-compute-2026-15), which closes at the end of 2026. The A6 monomer calculations are finished. The partner-molecule and interface calculations are planned to finish on Sigma before it closes. This proposal covers the work that remains after Sigma: the Paper 1 referee round (submission December 2026, reports expected February–April 2027), completion and revision of Paper 2 during 2027, and porting the protocol to Arrhenius.