Unified model access · A billing layer for agents

Turn model calls into a clearer route

One key, one bill, several protocols. Mengbi Router connects Claude, GPT, Gemini, and more so access, cost, and usage live in one workbench.

Manage your keys, models, and usage from the console.

API protocol compatibility

Keep your SDK. Change one line.

Choose the protocol that matches your existing tools instead of rebuilding your call layer.

OpenAI-compatible

For the OpenAI SDK, Codex, and most compatible clients.

/v1/chat/completions

Anthropic native

For Claude Code with native messages and streaming events.

/v1/messages

Gemini native

For Gemini CLI and apps using the native Gemini protocol.

/v1beta/models/*

Routing and billing

Built for long context and agent work

Model routing, caching, availability, and spend tracking meet at one boundary.

One key

Connect multiple model families and protocols.

Traceable spend

See model, tokens, and final charge per request.

Switchable routes

Choose a lane for speed or cost.

No silent downgrade

Keep requested and returned models verifiable.

Quick integration

Run your first request in three moves

  1. Create an account and key

    Create project keys in the console; keep the primary credential off clients.

  2. Choose a protocol and model

    Match your SDK to the OpenAI, Anthropic, or Gemini endpoint.

  3. Review usage and billing

    Reconcile tokens, discounts, and final charges at request level.

Current model pricing

Current model pricing

Compare input, output, and cache prices for every available model.

USD / 1M tokens
ModelInput (Official)Input (Mengbi)DiscountOutput (Official)Output (Mengbi)Discount
Claude Fable 5claude-fable-5$10.00Cache $1$4.29Cache $0.42857257% OFF$50.00$21.4357% OFF
Claude Opus 4.8claude-opus-4-8$5.00Cache $0.5$2.15Cache $0.21428657% OFF$25.00$10.7257% OFF
Claude Opus 5claude-opus-5$5.00Cache $0.5$2.15Cache $0.21428657% OFF$25.00$10.7257% OFF
Claude Sonnet 5claude-sonnet-5$2.00Cache $0.2$0.86Cache $0.08571557% OFF$10.00$4.2957% OFF
DeepSeek V4 Flash 福利版deepseek-v4-flash-free$0.44Cache $0.007$0.01Cache $0.0000199% OFF$1.32$0.0199% OFF
DeepSeek V4 Pro 福利版deepseek-v4-pro-free$1.32Cache $0.022$0.01Cache $0.00003299% OFF$3.96$0.0199% OFF
Gemini 3.1 Pro Previewgemini-3.1-pro-preview$2.00Cache $0.2$1.72Cache $0.17142914% OFF$12.00$10.2914% OFF
Gemini 3.6 Flashgemini-3.6-flash$0.75Cache $0.075$0.65Cache $0.06428614% OFF$3.75$3.2214% OFF
Gemini 3.7 Flashgemini-3.7-flash$0.75Cache $0.075$0.65Cache $0.06428614% OFF$3.75$3.2214% OFF
GPT-5.4gpt-5.4$2.50Cache $0.25$0.36Cache $0.03571585% OFF$15.00$2.1585% OFF
GPT-5.4-Minigpt-5.4-mini$0.75Cache $0.075$0.11Cache $0.01071585% OFF$4.50$0.6585% OFF
GPT-5.5gpt-5.5$5.00Cache $0.5$0.72Cache $0.07142985% OFF$30.00$4.2985% OFF
GPT-5.6 Lunagpt-5.6-luna$0.20Cache $0.02$0.03Cache $0.00285885% OFF$1.20$0.1885% OFF
GPT-5.6 Solgpt-5.6-sol$5.00Cache $0.5$0.72Cache $0.07142985% OFF$30.00$4.2985% OFF
GPT-5.6 Terragpt-5.6-terra$2.00Cache $0.2$0.29Cache $0.02857285% OFF$12.00$1.7285% OFF
OPENAI-COMPATIBLE / SERVER ENVIRONMENTKEY NEVER IN BROWSER
const client = new OpenAI({
  apiKey: process.env.MENGBI_ROUTER_KEY,
  baseURL: process.env.MENGBI_ROUTER_BASE_URL,
})

MENGbi / ROUTER

Give your next agent task a clearer route.

Start small, verify protocol, model, and cost, then move into production workflows.

Open console