Provider · verified 16 Oct 2026
Z.AI GLM 智谱
GLM-5.3 Flash on OpenRouter matches Z.AI’s published token price. GLM-5.3 the flagship does not: the OpenRouter list price is a cheap host. The Z.AI host on the same endpoints feed matches the docs at $1.40 input and $4.40 output per million tokens.
Models on the latest OpenRouter snapshot
| Model | Input | Output | vs Opus |
|---|---|---|---|
z-ai/glm-5.3-prime | $2.80 | $8.80 | 46% cheaper |
z-ai/glm-5.3-flashx | $0.370 | $1.25 | 93% cheaper |
z-ai/glm-5.3-flash | $0.024 | $1.13 | 96% cheaper |
z-ai/glm-5.3 | $0.079 | $4.80 | 84% cheaper |
z-ai/glm-5.2 | $0.620 | $2.42 | 87% cheaper |
z-ai/glm-5.1 | $0.966 | $3.04 | 81% cheaper |
z-ai/glm-5v-turbo | $1.20 | $4.00 | 76% cheaper |
z-ai/glm-5-turbo | $1.20 | $4.00 | 76% cheaper |
z-ai/glm-5 | $1.00 | $3.20 | 81% cheaper |
z-ai/glm-4.7-flash | $0.060 | $0.400 | 98% cheaper |
z-ai/glm-4.7 | $0.600 | $2.20 | 88% cheaper |
z-ai/glm-4.6v | $0.300 | $0.900 | 94% cheaper |
Guides
- This week in Chinese AI: prices checked, two list prices that lie
- Chinese AI API prices, live, versus the latest Opus and Sol
- Call Chinese models without sending the prompt to the lab
- Use Chinese models in Claude Code: the base URLs we could verify
- Coding plans compared: GLM's credits, and the plans we would not price
- What a coding session costs: flash models, flagships, and the cache assumption
- GLM Coding Plan: Lite, Pro, and Max, in credits not slogans
- Is it safe to use Chinese AI? A sourced risk matrix, not a slogan
- OpenRouter versus direct: where the Chinese model is actually cheaper
- Paying for Chinese AI from abroad: what the docs show, and what they do not
- Qwen, DeepSeek, GLM, Kimi: which Chinese model for which job
- Z.AI API: Flash, FlashX, and the flagship price people misread