Functional commitments (FCs) enable a prover to commit to a message and later produce a succinct proof of its image under any given admissible function. Unlike succinct non-interactive arguments (SNARGs), secure FCs can be realised under falsifiable assumptions in the standard model, making them attractive alternatives when fully algebraic constructions are desired. All known algebraic constructions of FC are either lattice-based or pairing-based. While a lattice-based FC for circuits with almost optimal complexities has been recently achieved [Wee, CRYPTO'25], state-of-the-art pairing-based FCs for bounded-width [Balbás-Catalano-Fiore-Lai (BCFL), TCC'23] and bounded-size circuits [Wee-Wu, EUROCRYPT'24] require prohibitive public parameter sizes and , respectively.
In this work, we present a new algebraic pairing-based FC which achieves public parameter size for bounded-width and unbounded depth circuits. The construction preserves all nice properties of BCFL: commitment size, proof size, additive homomorphism, efficient verification, and chainability. For bounded-size circuits, we alternatively obtain public parameters with proofs. At the core of our scheme lies a new chainable FC for quadratic functions with commitments computed with respect to a power basis, as well as techniques for switching between different commitment bases.