Z.ai: GLM 4.7 Flash

Z.ai's GLM-4.7-Flash, a 30B-class model created on January 19, 2026, is designed to balance performance and efficiency, particularly for agentic coding use cases. With a substantial context length of 200,000, it demonstrates strong reliability, achieving a 93% success rate across benchmarks. However, the model exhibits a significant weakness in speed, consistently showing longer response times, ranking in the 1st percentile. Its pricing is moderate, falling within the 38th percentile. In terms of performance across categories, GLM-4.7-Flash excels in Email Classification, achieving perfect 100% accuracy, making it the most accurate model at its price point and speed. It also shows strong capabilities in Coding (90.7% accuracy) and Ethics (98.0% accuracy). While its General Knowledge (96.5%) and Reasoning (78.0%) are respectable, its performance in Hallucinations (67.3%) and Mathematics (74.7%) is less competitive, ranking in the lower percentiles. The model's instruction following is average at 61.5%. Overall, GLM-4.7-Flash is a reliable model with notable strengths in classification and coding, but its slow processing speed and moderate hallucination and math scores are areas for potential improvement.