兔子API 兔子API
中

provider

Tencent Hunyuan

Direct answerFor direct TokenHub access, call /v1/models and use the case-sensitive model ID. Custom online inference uses a service ID. Guangzhou and international bases differ, while this gateway uses its own token and adapted route.

Updated · Reviewed

TokenHub versus Hunyuan

TokenHub is a Tencent Cloud hosting platform whose marketplace can list Hunyuan and third-party models, so not every entry belongs to Tencent Hunyuan. Verify author, capability, price, and lifecycle, especially pre-offline status. This gateway page declares only /v1/chat/completions; official TokenHub Responses, Anthropic, or other APIs are not gateway endpoints by implication.

Region, authentication, and this gateway

The Guangzhou service root is https://tokenhub.tencentmaas.com; international Singapore uses https://tokenhub-intl.tencentmaas.com, with OpenAI-compatible paths under /v1/.... Regions are not interchangeable. Direct calls use a TokenHub Bearer API key, while this gateway uses its own base URL and Bearer token. Legacy Hunyuan paths also cannot be spliced onto TokenHub.

Model ID versus service ID

For default models, call GET /v1/models and copy data[].id exactly, preserving case. A custom online inference deployment uses a service ID. Console display names are for humans. Through this gateway, use an enabled gateway name, which a channel may map to a TokenHub model ID or custom service ID.

Streaming, protocols, and errors

TokenHub officially supports OpenAI Chat, Responses, and Anthropic compatibility, but capability remains model-specific and gateway clients must follow this page's api_paths. Errors can include code, type, source, upstream detail, and request_id; 429 may include Retry-After. Retain the complete error to separate authentication, unsupported model/protocol, throttling, and upstream failures.

Cost and security

A TokenHub key can be scoped to default models or custom endpoints and assigned daily, monthly, and lifetime usage limits, but budgets must be actively enabled. Split keys by workload, cap concurrency and output, and alert on usage. Marketplace pricing, status, and geography can change, so recheck them before launch.

Reviewed provider catalog

Latest models by capability

These are provider-catalog models, not a promise of availability on this site. Confirm callable IDs, endpoints, and pricing in the model marketplace.

Text, reasoning, and code

  • hy3 Stable

    Tencent Hunyuan's production text flagship on TokenHub with a 256K context window.

    Origin: Tencent Hunyuan Release date not stated in the cited official source
    Official model catalog ↗ Model details →

Image generation and editing

  • hy-image-v3.0 Stable

    Tencent Hunyuan's 3.0 text-to-image and image-to-image model on TokenHub.

    Origin: Tencent Hunyuan Release date not stated in the cited official source
    Official model catalog ↗ Model details →

Video generation and editing

Embeddings, retrieval, and reranking

Check live availability in the model marketplace

Use cases

  • Chinese and enterprise tasks
  • Reasoning and multimodality
  • Tencent Cloud integration

API protocols

  • /v1/chat/completions

FAQ

Can I mix the Guangzhou and international TokenHub base URLs?

No. The Guangzhou service root is https://tokenhub.tencentmaas.com and international Singapore uses https://tokenhub-intl.tencentmaas.com, with OpenAI paths under /v1/.... Account, key, model, and region must match; gateway calls use this site's base URL and token.

Should model contain a model ID or an online-service ID?

For default models, call GET /v1/models and copy data[].id exactly, including case, while checking pre-offline status. A custom online inference deployment uses a service ID. A console display name replaces neither value.

How can I cap spending on a TokenHub API key?

A key can be scoped to default models or custom endpoints and can have daily, monthly, and lifetime usage limits, which may be disabled by default. Enable budgets and split keys by environment or workload to contain leakage.

How should a TokenHub 429 be retried?

Read code, type, source, request_id, and any Retry-After value to separate platform throttling from an upstream failure. Retry only transient cases according to Retry-After or bounded backoff and reduce concurrency; unsupported model or protocol errors need configuration changes.

  • OpenAI-compatible API guideThis site's OpenAI-compatible routes use a Bearer token; Responses, Chat Completions, and media endpoints have distinct objects and streaming semantics and must match the target model.
  • Tool use and AI agent model guideAn agent is a controlled loop, not merely tool calling: structured schemas, least privilege, state, idempotency, timeouts, approval, audit, and replayable evaluation.
  • Video generation and asynchronous API guideDesign video generation as an asynchronous job system with idempotent submit, persisted IDs, bounded polling or verified callbacks, terminal handling, output transfer, and moderation.
  • AI API quickstart: from API key to first requestAn AI API call needs a matching base URL, API key, model ID, and endpoint; start with minimal curl, then add error handling, cost controls, and key security.
  • Common AI API error codes and troubleshootingHandle AI API failures by parsing the HTTP status, provider code, error type, message, retry hints, and request ID together. Fix request, authentication, billing, and hard-quota causes first; retry only transient throttling, network failures, and selected 5xx errors with strict bounds.

Official sources

  1. Tencent Cloud TokenHub Model Marketplace Official
  2. Tencent Cloud TokenHub API Usage Official
  3. Tencent Hunyuan Migration Notice Official
  4. Tencent Cloud TokenHub Errors Official
  5. Tencent Cloud TokenHub API Key Management Official