Z.ai releases GLM-5.2, an MIT-licensed open-weight model reaching near-frontier performance on long-horizon agentic coding benchmarks

Z.ai has released GLM-5.2, an open-weight model (MIT license, no regional restrictions) with a 1M-token context window designed specifically for long-horizon agentic coding tasks. On FrontierSWE, PostTrainBench, and SWE-Marathon, it trails only Claude Opus 4.8 among all models tested, outperforming GPT-5.5 on FrontierSWE and ranking as the top open-source model across all three benchmarks. On standard coding benchmarks, GLM-5.2 closes to within a few points of Claude Opus 4.8 on Terminal-Bench 2.1 (81.0 vs. 85.0). For financial institutions, this release raises the structural question of whether open-weight models have reached sufficient capability parity with closed frontier models to serve as substitutes for enterprise agentic workloads, which would compress pricing power at the application and API layers.