BigModel.io documentation overview

BigModel.io is a global model API platform. Learn the OpenAI- and Anthropic-compatible APIs it serves, how keys, groups, models and USD balance fit together, and where to find each guide.

On this page
  1. Two protocols, one gateway
  2. Core concepts
  3. Map of the docs
  4. Check the service
  5. Troubleshooting
  6. Related

BigModel.io is a global model API platform: one API key for leading AI models, over the OpenAI and Anthropic protocols your tools already speak.

  • One key, two protocols. Use the same key with OpenAI SDKs and with Anthropic clients such as Claude Code.
  • Priced in US dollars. Model pages show the per-token price next to the official price.
  • Pay without an overseas card. Top up with Alipay or USDT, among the methods shown at checkout.
  • Built for coding agents. Guides for Claude Code, Codex CLI, Cursor and other tools.
  • Public status page. Availability and latency per model at /status.

Text models you can call today come from the GLM (Zhipu), DeepSeek, Kimi (Moonshot AI), MiniMax and Doubao (ByteDance) families. Image, video and audio models include Seedance, GPT Image, Qwen-Image, Kling, Vidu, Grok Imagine, Suno and more (see the multimodal overview). The model catalog always has the current list. Text models from other families, such as Claude, GPT, Gemini and Qwen, are not available yet; the catalog marks any that are coming soon.

You want to Start here
Send a first request in a few minutes API quickstart
Use a coding tool such as Claude Code or Cursor Claude Code, Cursor, Codex CLI
Move existing OpenRouter code Migrate from OpenRouter
Understand charges and top-ups Billing & top-ups
Decode an error response Errors

#Two protocols, one gateway

The same key works on both protocols, subject to what the key's group serves (see below).

Protocol Base URL to configure Main endpoints
OpenAI-compatible https://bigmodel.io/v1 POST /v1/chat/completions, POST /v1/responses, GET /v1/models
Anthropic-compatible https://bigmodel.io POST /v1/messages, POST /v1/messages/count_tokens (the client adds /v1)

Other endpoints: GET /v1/usage reports the key's balance and limits, image, video, avatar, speech and music models use /v1/images/*, /v1/videos and /v1/audio/*, and /v1/workflows/short-drama turns a script into clips (see the multimodal overview).

bash
curl https://bigmodel.io/v1/chat/completions \
  -H "Authorization: Bearer $BIGMODEL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"glm-5.3","messages":[{"role":"user","content":"Hello!"}]}'

#Core concepts

API key. Created in the console under API Keys. A key looks like sk- followed by 64 hex characters. You can view and copy the full key there at any time. Each key can have its own quota, spend windows, expiry and IP rules. See Authentication.

Group. Every key belongs to one group. The group decides which models and endpoints the key can use and which rate multiplier applies to its charges. Some groups (for example GLM or DeepSeek groups) serve only /v1/chat/completions; their keys get a 404 on /v1/messages and /v1/responses.

Model ID. Plain lowercase IDs that keep their dots, such as glm-5.3, kimi-k3 or deepseek-v4-pro. Copy them from the model catalog. Availability depends on the key's group; GET /v1/models lists what your key can call.

Balance. Your account holds a USD balance. Each request is charged by tokens at the model's price times the group multiplier. Paid models need a balance above zero; models priced at $0 can be called at a zero balance. New accounts get free trial credit; the amount is shown in the console. Plans (subscriptions) are an alternative for keys in a plan group. See Billing & top-ups.

Usage. Every request appears in Usage with tokens, cost, model and timing. GET /v1/usage returns the same totals for one key.

#Map of the docs

#Check the service

The status page shows availability, latency and time to first token per model. Questions about your account go to Support or support@bigmodel.io.

#Troubleshooting

Not sure which base URL a tool wants. If it asks for a base URL on the OpenAI protocol, use https://bigmodel.io/v1; Anthropic clients take https://bigmodel.io. See the quickstart.

A model works in the catalog but not with your key. The key's group does not serve it, and the request returns 503 No available accounts. See Errors.

Updated · View as Markdown