"GLM-5 beats Claude Opus on the world’s hardest benchmark" This was true a couple weeks ago. The Z.ai models are going haywire recently. It might be bad tech, it might be malicious external actors, it might be both.…
"GLM-5 beats Claude Opus on the world’s hardest benchmark" This was true a couple weeks ago. The Z.ai models are going haywire recently. It might be bad tech, it might be malicious external actors, it might be both.…