GLM-5.2 vs Opus comparison thread

Category: Pulse
Status: discussion
Sources: comparison post, discussion
Summary: A head-to-head post pitting GLM-5.2 against Claude Opus 4.8 on a one-shot WebGL platformer prompt drew a large HN thread (about 480 points).
Comments: Top commenters pushed back that a single one-shot prompt is not a benchmark and not representative of collaborative agent use; some said price comparisons should include Claude subscription tiers, not only pay-as-you-go output cost; practitioners reported GLM-5.2 is slow to start and strays during planning but corrects, while one called it the most interesting open-weight release of the year and another said Chinese models over-optimize for benchmarks.