π©
GLM-5.2 & GLM-5.1 are flagship models designed for agentic engineering and long-horizon tasks with advanced coding capabilities. GLM-5.2 features a solid 1M-token context and achieves state-of-the-art performance on coding benchmarks (81.0 on Terminal-Bench 2.1, 62.1 on SWE-bench Pro), approaching closed-source frontier models while remaining the strongest open-source option.
