The cheapest model on our gateway and one of the four serving right now. Copy the code below and it works on the first try.
Get API Key — $1 Free Credit See All Live ModelsGLM-4 Flash is Zhipu AI's (智谱AI) small, fast model, and at $0.10 per million input tokens it is the lowest-cost option we serve. It is live on this gateway today — unlike the larger GLM-5 and GLM-5.1 models, whose upstream account is currently suspended and which return HTTP 503.
The cheapest model on the gateway. Good for high-volume classification, extraction and routing.
Not a roadmap item. It answers requests today; the code sample below is tested against the live endpoint.
Same request shape as OpenAI. Point the base URL at us and change one string.
Function calling, JSON mode and streaming all supported.
| Metric | TokenEase Price |
|---|---|
| Input tokens | $0.10 / million |
| Output tokens | $0.40 / million |
| Context window | 128,000 tokens |
| Free trial | $1 credit on signup |
from openai import OpenAI
client = OpenAI(
base_url="https://tokenease.io/v1",
api_key="sk-your-tokenease-key"
)
response = client.chat.completions.create(
model="glm-4-flash",
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)
Four models serve today: glm-4-flash, kimi-k3 (256K context, flagship), kimi-k2.7-code and kimi-k2.7-code-highspeed. Everything else on the roster is paused upstream. The authoritative list is the API itself:
curl -s https://tokenease.io/v1/models
Yes. glm-4 is a separate, larger model whose upstream account has an unpaid balance and currently returns 503. glm-4-flash is the fast variant and is serving normally.
No. Direct Zhipu access needs a Chinese account. TokenEase gives you a key on email signup.
Call /v1/models. Every entry carries an enabled flag and, when false, a disabled_reason.