ChinaGenAIFree API key
Compare API accessAll guidesWhat’s new
Search

Tencent TokenHub Free API Trial: Connect the OpenAI-Compatible Endpoint

By ChinaGenAI · Updated

Set the regional TokenHub base URL, list models available to your key, then check the trial quota and billing boundary before connecting an app.

Tencent Cloud's LLM Service TokenHub provides an OpenAI-compatible API for Tencent HY and other model families, including DeepSeek, GLM, Kimi, MiniMax and MiMo. This is a provider account and API route; it is separate from Tencent's consumer chat apps. The steps below use Tencent's published documentation, not an account or live inference test.

What “free trial” means here

Tencent documents a one-time new-user trial through 31 December 2026. Its current trial page lists 1 million free tokens for language models with a one-year validity period. The same page says models under one root account share the quota and a package can be claimed only once per account per model. Treat it as a shared account allowance; check the balance shown for your model instead of assuming every model adds a separate 1 million tokens.

There is also a USD 10 account freeze each time you claim a model's free package or enable postpaid billing. Tencent says the amount is released after the model resources are terminated; deleting the default endpoint requires customer support. If free quota runs out and postpaid is off, calls stop. If you enable postpaid, further usage is billed by token.

Tencent: trial quota, expiry and frozen amount

Match the endpoint to the account region

Tencent's English API instructions list these OpenAI-compatible regional base URLs. Use the key created for the same site and region; Tencent says regional sites and keys are not interchangeable.

  • Singapore, global: https://tokenhub-intl.tencentcloudmaas.com/v1
  • Guangzhou, Chinese mainland: https://tokenhub.tencentcloudmaas.com/v1
  • Silicon Valley, United States: https://tokenhub-us.tencentcloudmaas.com/v1

One detail to verify before copying: Tencent's Chinese API page currently shows different hostnames ending in tencentmaas.com, while its English API instructions show tencentcloudmaas.com. Follow the current regional API instructions for the site where you created the key; do not substitute a hostname just because another tutorial uses it.

For the English Singapore example, set the base URL and provide your key through your local environment or secret manager:

Set the Tencent TokenHub global API base URL
export TOKENHUB_BASE_URL="https://tokenhub-intl.tencentcloudmaas.com/v1"

Find the model ID your key can use

TokenHub's authenticated GET /v1/models endpoint returns model IDs and availability status for the selected account. Use the ID from that response in your request. The list can change, and catalog names shown in older tutorials may not match your account.

List models available to your TokenHub key
curl "$TOKENHUB_BASE_URL/models" \
  -H "Authorization: Bearer $TOKENHUB_API_KEY"

Replace MODEL_ID_FROM_YOUR_MODELS_LIST below with an exact ID from that response. The base URL already ends in /v1; append only /chat/completions.

Send a non-streaming OpenAI-compatible Chat Completions request
curl "$TOKENHUB_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $TOKENHUB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID_FROM_YOUR_MODELS_LIST",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}],
    "stream": false
  }'

For a client that automatically builds Chat Completions URLs, configure its base URL as the regional address ending in /v1, not the full /chat/completions URL.

Check the common setup errors

  • 401 or 401002: compare the key's region/site with the hostname, include the Bearer header, and check the exact key used for /v1/models.
  • 402: check that the model is activated for inference and that trial quota remains. Do not enable postpaid unless you intend to use token-billed service.
  • 429: check the current rate limit and any Retry-After response header before retrying.
  • Model not found: copy the exact id from the authenticated model list; the display name and API parameter can differ.

API hosts, Bearer authentication and GET /v1/models · Model protocols and error handling · Quick-start examples for Python, Node.js, Java and Go

Compare it with other routes

TokenHub is one provider-hosted option, not a guarantee of unlimited free inference. Compare its trial conditions with other current providers in the API route comparison, or browse the Chinese-model API guide. ChinaGenAI's own free route is separate and selects its upstream model automatically; it does not promise a fixed Tencent or other model ID.

Sources and review scope

Checked 2 November 2026. Trial eligibility, balances, endpoint availability and inference were not tested. The documented free offer is time-limited; check Tencent's current console terms before connecting a client.

Free API access and limits

Check the site free API access and limits before trying to obtain a site key. Read the current status and request limits before using a site key. It is not a source of official provider keys.

ChinaGenAI is not affiliated with the providers named here. This release contains setup guidance and explicitly labeled fictional examples, not claimed live model benchmarks.