
Formerly Zhipu AI: the GLM family of open frontier models, strong at coding and agentic tasks.
Compare Fast and Cheap modes, see the percentage saved, or filter to models with a verified discount. Select any model for its full pricing and API options.
Fast mode prioritizes response speed. Cheap mode uses lower-cost capacity and may be a little slower. Compare the same model’s input and output rates below.
| Model | Fast mode Input / output per 1M tokens | Cheap mode Input / output per 1M tokens | Discount | Run model |
|---|---|---|---|---|
| GLM 5.3 FlashX LegalZ.ai · 1.05M context | $0.3885 / $1.3125 | No verified discount | — | |
| GLM 5.3 FlashXZ.ai · 1.05M context | $0.3885 / $1.3125 | No verified discount | — | |
| GLM 5.3 FlashZ.ai · 1.05M context | $0.0352 / $0.0872 | No verified discount | — | |
| GLM 5.3 Flash (batch)Z.ai · 1.05M context | $0.0525 / $0.189 | No verified discount | — | |
| GLM 5.3Z.ai · 1.05M context | $0.0734 / $1.47 | No verified discount | — | |
| GLM 5.3 (batch)Z.ai · 1.05M context | $0.3297 / $1.3346 | No verified discount | — | |
| GLM 5.2Z.ai · 1.05M context | $0.0063 / $7.14 | No verified discount | — | |
| GLM 5.2 (batch)Z.ai · 1.05M context | $0.3297 / $1.3346 | No verified discount | — | |
| GLM 5.1Z.ai · 205K context | $1.47 / $4.62 | $0.9437 / $2.966 | 35.8% off | |
| GLM 5V TurboZ.ai · 203K context | $1.26 / $4.2 | No verified discount | — | |
| GLM 5 TurboZ.ai · 203K context | $1.26 / $4.2 | No verified discount | — | |
| GLM 5Z.ai · 205K context | $1.05 / $3.36 | $0.63 / $2.016 | 40% off | |
| GLM 4.7 FlashZ.ai · 200K context | $0.0635 / $0.42 | No verified discount | — | |
| GLM 4.7Z.ai · 205K context | $0.63 / $2.31 | $0.336 / $1.47 | 36.4–46.7% off | |
| GLM 4.6Z.ai · 205K context | $0.525 / $2.1 | $0.42 / $1.68 | 20% off | |
| GLM 4.5Z.ai · 131K context | $0.63 / $2.31 | No verified discount | — | |
| GLM 4.5 AirZ.ai · 131K context | $0.1365 / $0.8925 | No verified discount | — |
Discounts compare current uncached input and output rates. Different percentages mean savings depend on your token mix. Fast rates are routing ceilings; actual charges and capacity may vary. No discount is shown without a verified quote.