GLM-4.7-Flash is a 30B-A3B MoE model optimized for speed and cost-efficiency. It achieves SWE-bench 59.2%, TAU-Bench 79.5%, and AIME 91.6%, making it ideal for coding and reasoning tasks. With 128K context and 131K output limits, it offers exceptional value at $0.07/$0.40 per million tokens, 10x cheaper than standard GLM-4.7. Best for developers prioritizing speed and cost over maximum capability.