We investigate the efficiency of a (module-)LWR-based PRF built using the GGM design. Our construction enjoys the security proof of the GGM construction and the (module-)LWR hardness assumption which is believed to be post-quantum secure. We propose GGM-based PRFs from PRGs with larger ratio of output to input. This reduces the number of PRG invocations which improves the PRF performance and reduces the security loss in the GGM security reduction. Our construction bridges the gap between practical and provably secure PRFs. We demonstrate the efficiency of our construction by providing parameters achieving at least 128-bit post-quantum security and optimized implementations utilizing AVX2 vector instructions. Our PRF requires, on average, only 39.4 cycles per output byte.