# Cheapest GLM 5.2 API providers, compared (prices checked Nov 7, 2026)

URL: https://ernd.ai/inference-prices/glm-5-2

> ernd partners with [Wafer](https://ernd.ai/partners/wafer). Prices are from OpenRouter's public listings.

GLM 5.2 is an open-weight model from Z.ai. On OpenRouter, 33 listings from 26 providers sell it, and the blended price runs from $0.29 to $3.66 per million tokens.

| Provider | Input /1M | Output /1M | Blended 3:1 | Cache read /1M | Context | Quantization | Latency | Throughput | With ernd card |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Baidu Qianfan (fp8) | $0.189 | $0.594 | $0.29 | $0.0351 | 1.05M | FP8 | 0.95 s | 65 tok/s | – |
| StreamLake | $0.28 | $0.88 | $0.43 | $0.052 | 1.02M | FP8 | 1.50 s | 45 tok/s | – |
| DeepInfra (Flex) | $0.45 | $1.44 | $0.698 | $0.084 | 1.05M | FP4 | 3.04 s | 37 tok/s | – |
| DeepInfra | $0.563 | $1.80 | $0.872 | $0.105 | 1.05M | FP4 | 1.40 s | 40 tok/s | – |
| Ambient | $0.60 | $2.00 | $0.95 | $0.15 | 203K | FP8 | 0.99 s | 76 tok/s | – |
| NovitaAI | $0.65 | $2.04 | $0.998 | $0.121 | 1.05M | FP8 | 2.78 s | 40 tok/s | – |
| CoreWeave | $0.62 | $2.42 | $1.07 | $0.14 | 1.05M | FP4 | 0.54 s | 145 tok/s | – |
| DigitalOcean | $0.70 | $2.20 | $1.07 | $0.105 | 1.05M | Not listed | 0.97 s | 73 tok/s | – |
| SiliconFlow | $0.70 | $2.20 | $1.07 | $0.13 | 1.05M | FP8 | 1.08 s | 80 tok/s | – |
| Morph | $0.74 | $3.00 | $1.30 | $0.25 | 1.05M | FP8 | 4.64 s | 32 tok/s | – |
| AtlasCloud | $0.938 | $2.95 | $1.44 | $0.174 | 1.05M | FP8 | 1.22 s | 50 tok/s | – |
| Alibaba Cloud Int. | $0.966 | $3.04 | $1.48 | $0.193 | 1M | FP8 | 1.02 s | 67 tok/s | – |
| Wafer (US), ernd partner | $0.006 | $6.80 | $1.70 | $0.005 | 1.05M | FP4 | 0.46 s | 116 tok/s | $1.53 after 10% cash back |
| Alibaba Cloud Int. (US-EAST-1) | $1.12 | $3.52 | $1.72 | $0.224 | 1M | FP8 | 1.50 s | 54 tok/s | – |
| Relace | $0.0065 | $7.91 | $1.98 | $0.0065 | 1.05M | FP4 | 1.37 s | 49 tok/s | – |
| Cloudflare | $0.19 | $8.00 | $2.14 | $0.19 | 262K | Not listed | 0.98 s | 85 tok/s | – |
| Wafer, ernd partner | $0.19 | $8.00 | $2.14 | $0.18 | 1.05M | FP4 | 0.42 s | 104 tok/s | $1.93 after 10% cash back |
| Baseten | $1.40 | $4.40 | $2.15 | $0.14 | 1.05M | FP8 | 2.41 s | 51 tok/s | – |
| Baseten (US) | $1.40 | $4.40 | $2.15 | $0.14 | 1.05M | FP8 | 2.52 s | 55 tok/s | – |
| Friendli | $1.40 | $4.40 | $2.15 | $0.26 | 1.05M | Not listed | 0.43 s | 93 tok/s | – |
| GMICloud | $1.40 | $4.40 | $2.15 | $0.26 | 1.05M | FP8 | 2.88 s | 33 tok/s | – |
| Inceptron | $1.40 | $4.40 | $2.15 | $0.19 | 1.05M | FP4 | 0.53 s | 106 tok/s | – |
| Mistral (nvfp4, ZDR) | $1.40 | $4.40 | $2.15 | $0.14 | 1.05M | NVFP4 | 1.27 s | 77 tok/s | – |
| Nebius Token Factory | $1.40 | $4.40 | $2.15 | $0.14 | 1.05M | FP4 | 1.76 s | 45 tok/s | – |
| Parasail | $1.40 | $4.40 | $2.15 | $0.26 | 262K | FP4 | 0.87 s | 105 tok/s | – |
| Together | $1.40 | $4.40 | $2.15 | $0.26 | 1.05M | Not listed | 0.36 s | 101 tok/s | – |
| Venice | $1.40 | $4.40 | $2.15 | $0.26 | 1M | FP8 | 2.39 s | 78 tok/s | – |
| Z.ai | $1.40 | $4.40 | $2.15 | $0.26 | 1.05M | FP8 | 1.42 s | 54 tok/s | – |
| Mistral (EU) | $1.54 | $4.84 | $2.37 | $0.154 | 1.05M | Not listed | 0.79 s | 97 tok/s | – |
| Baseten (Fast) | $2.10 | $6.60 | $3.23 | $0.21 | 1.05M | FP8 | 0.39 s | 194 tok/s | – |
| Baseten (US, Fast) | $2.10 | $6.60 | $3.23 | $0.21 | 1.05M | FP8 | 0.40 s | 194 tok/s | – |
| Alibaba Cloud Int. (Fast) | $2.31 | $7.26 | $3.55 | $0.462 | 1M | FP8 | 1.05 s | 79 tok/s | – |
| Baidu Qianfan (fp4, Fast) | $2.25 | $7.88 | $3.66 | $0.56 | 1.05M | FP4 | 1.02 s | 69 tok/s | – |

Blended is three parts input price to one part output, per million tokens. Latency is time to first token and throughput is output tokens per second, both averaged from OpenRouter's daily figures for the last 7 days. Cheapest and fastest only count listings that were up when checked.

Wafer is an ernd partner: 10% back on Wafer inference with the ernd card. With ernd card is the blended price less 10% cash back on Wafer charges. It comes back on the card bill and doesn't change Wafer's list price. It applies when you pay Wafer directly, not through OpenRouter credits.

## Questions

### What is the cheapest GLM 5.2 API provider?

Baidu Qianfan (fp8), at $0.29 per million tokens blended ($0.189 input, $0.594 output), in OpenRouter's listings on November 7, 2026.

### How much does GLM 5.2 cost per million tokens?

Input runs from $0.006 to $2.31 per million tokens across 33 listings, and output from $0.594 to $8.00. The middle listing comes to $2.14 blended.

### Which GLM 5.2 provider is fastest?

Baseten (Fast) had the most throughput over the last 7 days, about 194 tokens per second. Together was quickest to the first token at 0.36 s.

### How do I pay for GLM 5.2 inference?

Buy OpenRouter credits and it sends each request to one of these providers, or open an account with a provider and pay it directly. The ernd card is a credit card for US companies. It earns 4% back on AI vendor spend and 1.5% back on everything else. Wafer charges on the card earn 10% back.

Source: OpenRouter public model and provider listings (https://openrouter.ai/models), checked November 7, 2026.

The ernd card is a credit card for US companies. It earns 4% back on AI vendor spend and 1.5% back on everything else. [Request the card at app.ernd.ai](https://app.ernd.ai/login?next=%2Fapp%2Fcard%2Fapply&utm_source=llms&utm_campaign=inference_prices).