GLM-5.2: Tops Multiple AI Benchmarks

GLM-5.2 seized multiple #1 rankings among open-weight models while outperforming several frontier systems, according to data shared by analyst @milesdeutscher on June 23, 2026. The model posted 1360 Elo in Design Arena to edge Fable 5, crossed 80% on Terminal-Bench 2.1 with a best run of 82.7%, and led the Artificial Analysis Intelligence Index v4.1 at a score of 51; it also stayed competitive or ahead of GPT-5.5 on GDPval-AA v2 and LiveBench Agentic Coding. Additional top-three finishes include 74.4% on FrontierSWE and 62.1% on SWE-bench Pro, where it surpassed GPT-5.5's 58.6%, underscoring rapid gains in open-weight model benchmarks and frontier model comparisons.