GLM‑5.2 matches Claude Opus in real‑world coding for under $4
Lenny tested Z.AI’s open-weight GLM-5.2 on four production coding tasks – a code‑base audit, UI redesign, and a 45‑minute autonomous bug‑hunting run – and found it on par with Claude Opus while spending only $3.36 for ~6 M tokens. The cheap, vendor‑independent model proves viable for cost‑conscious teams.