GLM-5.2 Turbo

A speed-optimized, hosted "Turbo" serving tier of Z.ai's GLM-5.2, surfaced Aug 17 2026 (API id glm-5.2-fast). It targets latency-sensitive coding and agent workloads at the premium fast tier and carries GLM-5.2's 1M-token context. Served through Z.ai and SCX.ai with list pricing around $1.99 input / $6.16 output per Mtok — well above standard GLM-5.2 ($0.55/$1.78), reflecting the dedicated fast-serving tier. Z.ai has not published a separate parameter count, architecture detail, or open-weight release for the Turbo variant, so size and weights are recorded undisclosed pending confirmation; the underlying GLM-5.2 base is an open-weight (MIT) ~753B MoE. Added this sweep to resolve a catalog gap prior runs flagged for lacking a solid spec source (now sourced from the LLM Gateway model page); fields beyond context, pricing, and providers are conservative and unverified.