Z.ai: GLM 5.2.
Open-weights text model from Z.ai for long-horizon coding and agent work, with a one million token context and adjustable thinking effort, under the MIT licence.
Cheapest hosted endpoints.
Variants in the GLM family.
Z.ai: GLM 5.3 Flash
Open-weights multimodal model from Z.ai with 320B total and 18B active parame...
glm-5.3
GLM-5.3 uses the same base model as GLM-5.2 — every gain comes from post-trai...
GLM-4.5
Zhipu's frontier open-weight MoE — 355B total, 32B active. Strong agentic + r...
GLM-4.5-Air
Smaller, cheaper sibling of GLM-4.5. 106B total, 12B active.
GLM-5.1
Zhipu's GLM 5.1 series — successor to GLM-5 on z.ai's API.
GLM-5
Zhipu's GLM 5 generation — closed flagship between GLM-4.7 and GLM-5.1.
GLM-5 Turbo
Faster, cheaper sibling of GLM-5 on z.ai.
GLM-4.7
Open-weights coding model from Z.ai for English and Chinese, improving on GLM...
GLM-4.6
Open-weights English and Chinese text model from Z.ai with a 200K token conte...
GLM-4.5-Flash
Lowest-latency, lowest-cost variant of GLM-4.5 on z.ai.
Z.ai: GLM 5.3 FlashX
GLM-5.3-FlashX is the high-speed variant of Z.ai's GLM-5.3-Flash, a native mu...
Z.ai: GLM 5.3 FlashX Legal
GLM-5.3-FlashX-Legal (GLM-Legal) is Z.ai's legal-domain variant of GLM-5.3-Fl...
Z.ai: GLM 4 32B
GLM 4 32B is a cost-effective foundation language model. It can efficiently p...
Z.ai: GLM 4.7 Flash
As a 30B-class SOTA model, GLM-4.7-Flash offers a new option that balances pe...
Z.ai: GLM 5V Turbo
GLM-5V-Turbo is Z.ai’s first native multimodal agent foundation model, built ...
What it costs per month across providers.
Estimate your monthly bill for Z.ai: GLM 5.2 across every host that publishes per-token pricing. Slide your token volumes; the chart + table re-rank cheapest-first.
Cheapest provider on the left.
Total monthly cost — input + output tokens combined.
Bill breakdown.
About Z.ai: GLM 5.2.
A Z.ai model for long-horizon tasks, released with open weights under the MIT licence for English and Chinese text generation. It steps up from GLM-5.1 by sustaining work over a one million token context. It offers several thinking effort levels to trade quality against latency, and an indexer-sharing design that cuts per-token compute at that context length. Its multi-token prediction layer also helps speculative decoding.