GLM-5.2
Z.aiOpen weights1.05M context
GLM-5.2 is an open-weights model from Z.ai with a 1.05M-token context window. 17 providers list it, from $0.75 per million input tokens and $2.40 per million output tokens (DeepInfra).
- Maker
- Z.ai
- Licence
- Open weights
- Context window
- 1.05M
- Providers listed
- 17
- Lowest input, per 1M
- $0.75
- Lowest output, per 1M
- $2.40
- Effort settings
- Non-reasoning, Low, Xhigh, Max
Providers serving GLM-5.2
Cheapest first, by the blended rate of three input tokens to one output token. Prices in US dollars per million tokens, as of 26 Oct 2026.
| Provider | Input | Output | Context | Tools | JSON mode | OpenAI-compatible | Model ID |
|---|---|---|---|---|---|---|---|
DeepInfra | $0.75 | $2.40 | 1.05M | Yes | Yes | No | zai-org/GLM-5.2 |
SiliconFlow | $1.30 | $4.09 | 1.05M | Yes | Yes | Yes | zai-org/GLM-5.2 |
Alibaba Cloud | $1.40 | $4.40 | 1M | Yes | Yes | Yes | glm-5.2 |
Arcee AI | $1.40 | $4.40 | 262k | Yes | Yes | Yes | zai-org/glm-5.2 |
Baseten | $1.40 | $4.40 | 1.05M | Yes | Yes | Yes | zai-org/GLM-5.2 |
Cloudflare | $1.40 | $4.40 | 262k | Yes | Yes | Yes | @cf/zai-org/glm-5.2 |
Databricks Billed in DBUs; converted at Databricks' $0.07 per DBU (AWS). | $1.40 | $4.40 | 1M | Yes | Yes | Yes | databricks-glm-5-2 |
DigitalOcean | $1.40 | $4.40 | 262k | Yes | Yes | Yes | glm-5.2 |
FriendliAI | $1.40 | $4.40 | 1.05M | Yes | Yes | Yes | zai-org/GLM-5.2 |
GMI (FP8) | $1.40 | $4.40 | 1M | Yes | Yes | Yes | zai-org/GLM-5.2-FP8 |
Google Vertex | $1.40 | $4.40 | 1M | Yes | Yes | No | zai-org/glm-5.2-maas |
Mistral | $1.40 | $4.40 | 1.05M | Yes | Yes | No | glm-5-2 |
Nebius | $1.40 | $4.40 | 1.05M | Yes | Yes | Yes | zai-org/GLM-5.2 |
Novita | $1.40 | $4.40 | 1.05M | Yes | Yes | Yes | zai-org/glm-5.2 |
Together AI | $1.40 | $4.40 | 1.05M | Yes | Yes | No | zai-org/GLM-5.2 |
Wafer | $1.40 | $4.40 | 1.05M | Yes | Yes | Yes | GLM-5.2 |
Scaleway Priced in EUR only: €1.80 in / €5.50 out per 1M tokens (not converted) | — | — | 256k | Yes | Yes | Yes | glm-5.2 |
Benchmark scores
Published scores only, each linked to its source. Compare every model in Bench rankings