GLM-5.3 Flash launches, rivals Opus 4.8 at fraction of cost

Z.ai's GLM-5.3 Flash (formerly Ox Alpha) packs 320B parameters, 18B active, 1M context, and hybrid attention. On DeepSWE it nearly matches Luna while doing twice the work for the same budget, and rivals Opus 4.8 on coding benchmarks per Unsloth testing.