“Completely tucked into pines, yet the whole east wall dissolves into glass over a lake — a fire going, a boat at the dock. That tension between hiding and being wide open is where I’d be happiest.”
514s to generate40,877 output tokens
← All AI models
Z.ai · GLM
GLM 5.2
Generally availableText / reasoning model
Z.ai’s open-weights flagship (Zhipu) — strong long-horizon coding at a fraction of frontier cost.
Yes — GLM 5.2 is a real Z.ai model.
Released 2026 by Z.ai, GLM 5.2 is part of the GLM family. It's available through Z.ai's API, and we run it on PromptFrenzy — the examples below are real, unedited generations it produced.
Maker
Z.aiReleased
2026
Input price
$1 / 1M
Output price
$3.2 / 1M
What is GLM 5.2?
GLM 5.2 is Z.ai’s (Zhipu) open-weights flagship, known for strong long-horizon coding at a fraction of the cost of closed frontier models. We ran it in our Dream Home voxel showdown by reader request, direct on the z.ai API with thinking enabled, under the identical one-shot brief as the other models. It is an editorial entrant only: with reasoning on it exceeds our 300-second live-generation ceiling, so we don’t offer it as a runnable lane on the site.
Strengths
- ✓Open-weights — self-hostable, unusual for a frontier-class model.
- ✓Strong at long-horizon coding and structured generation.
- ✓Far cheaper per token than closed frontier models.
Trade-offs
- ✗Reasoning-on latency exceeds our 300s live ceiling — editorial-only here, not a runnable lane.
- ✗The z.ai API exposes no reasoning-effort control (thinking is on/off only).
What GLM 5.2 actually produces
Real, unedited outputs from GLM 5.2 in our model showdowns — same brief, same constraints as every other model, one shot each.
See GLM 5.2 compared head-to-head
Compare every model side-by-side
See how GLM 5.2 stacks up against the other frontier models on the same brief.
Browse all model showdowns →