Z.ai: GLM 4.6

Z.ai: GLM 4.6, released on September 30, 2025, demonstrates a significant expansion in context length to 200K tokens, enabling more complex agentic tasks. While the model tends to have longer response times, ranking in the 12th percentile for speed, and is positioned at premium pricing levels (14th percentile), it exhibits exceptional reliability with a 97% success rate across benchmarks. GLM 4.6 showcases strong performance in several key areas. It achieves perfect accuracy in both General Knowledge and Email Classification, with the latter also being the most accurate and fastest among models at its price point. Its coding performance is a notable strength, scoring 95.0% accuracy (92nd percentile) and demonstrating real-world improvements in various coding applications. Mathematics also stands out with 96.0% accuracy (93rd percentile). Reasoning performance is robust at 93.9% accuracy (78th percentile), supporting its claim of advanced reasoning and tool use. However, the model shows weaknesses in Instruction Following, with a low 3.1% accuracy (18th percentile), indicating challenges with complex multi-step directives. Its Ethics performance is also relatively low at 82.0% accuracy (18th percentile). Hallucinations are well-managed at 96.0% accuracy, suggesting a good ability to acknowledge uncertainty. Overall, GLM 4.6 excels in knowledge-based and coding tasks, but users should be mindful of its instruction following and ethical reasoning capabilities, as well as its slower response times and premium cost.