GLM 5.1 crushes every other model except Opus in agentic benchmark at about 1/3 of the Opus cost
https://preview.redd.it/s9lg647zjeug1.png?width=1161&format=png&auto=webp&s=4d0c361b5fbee97e4084e2d48543cafbc299ce25 I want to know whether GLM is another benchmark optimized model or actually useful in agents like OpenClaw, so I test…