# Baidu AI Cloud — LLM Models on FluxA Baidu Qianfan model gateway. OpenAI-compatible API across DeepSeek, ERNIE, GLM, Kimi, MiniMax and more. ## Models & Pricing Rates in **Units per 1M tokens**. 1 USDC buys 100,000 Units, and 1 MC buys 100,000 Units. See https://agentmarket.fluxapay.xyz/currency.md | Model ID | Display | Categories | Interfaces | Input (Units/1M) | Output (Units/1M) | |----------|---------|------------|------------|------------------|-------------------| | anthropic/claude-opus-5-5 | Claude Opus 5.5 | text, vision, thinking | messages | 288,000 | 1,440,000 | | anthropic/claude-fable-5 | Claude Fable 5 | text, vision, thinking | messages | 720,000 | 3,600,000 | | anthropic/claude-opus-5 | Claude Opus 5 | text, vision, thinking | messages | 340,000 | 1,700,000 | | anthropic/claude-opus-4-8 | Claude Opus 4.8 | text, vision, thinking | messages | 340,000 | 1,700,000 | | anthropic/claude-opus-4-7 | Claude Opus 4.7 | text, vision, thinking | messages | 340,000 | 1,700,000 | | anthropic/claude-opus-4-6 | Claude Opus 4.6 | text, vision, thinking | messages | 340,000 | 1,700,000 | | anthropic/claude-sonnet-5 | Claude Sonnet 5 | text, vision, thinking | messages | 136,000 | 680,000 | | anthropic/claude-sonnet-4-6 | Claude Sonnet 4.6 | text, vision, thinking | messages | 204,000 | 1,020,000 | | anthropic/claude-haiku-4-5-20251001 | Claude Haiku 4.5 | text, vision, thinking | messages | 68,000 | 340,000 | | openai/gpt-6-astra | GPT 6 Astra | text, vision, thinking | chat | 680,000 | 3,400,000 | | openai/gpt-6-sol | GPT 6 Sol | text, vision, thinking | chat | 136,000 | 680,000 | | openai/gpt-6-luna | GPT 6 Luna | text, vision, thinking | chat | 6,800 | 34,000 | | openai/gpt-5.6-sol | GPT 5.6 Sol | text, vision, thinking | chat | 272,000 | 1,360,000 | | openai/gpt-5.6-terra | GPT 5.6 Terra | text, vision, thinking | chat | 136,000 | 816,000 | | openai/gpt-5.6-luna | GPT 5.6 Luna | text, vision, thinking | chat | 13,600 | 81,600 | | deepseek/deepseek-v4-pro | Deepseek V4 Pro | text, thinking | chat | 118,000 | 237,000 | | deepseek/deepseek-v3.2 | DeepSeek V3.2 | text, thinking | chat | 20,000 | 29,000 | | moonshotai/kimi-k3 | Kimi K3 | text, vision, thinking | chat | 192,000 | 962,000 | | moonshotai/kimi-k2.7-code | Kimi K2.7 Code | text, vision, thinking | chat | 96,000 | 399,000 | | moonshotai/kimi-k2.6 | Kimi K2.6 | text | chat | 67,000 | 280,000 | | z-ai/glm-5.3 | GLM 5.3 | text, thinking | chat | 59,200 | 207,000 | | z-ai/glm-5.3-flash | GLM 5.3 Flash | text, thinking | chat | 5,900 | 20,700 | | z-ai/glm-5 | GLM-5 | thinking | chat | 70,000 | 224,000 | | z-ai/glm-5.1 | GLM 5.1 | text | chat | 98,000 | 308,000 | | bytedance-seed/doubao-seed-2-0-mini-260215 | Doubao Seed 2.0 Mini | text, thinking | chat | 3,000 | 29,600 | ## Calling Models OpenAI Chat Completions wire format, for every model whose Interfaces include `chat`. Drop-in for any OpenAI client, just swap the `baseURL`: `anthropic/claude-opus-5-5`, `anthropic/claude-fable-5`, `anthropic/claude-opus-5`, `anthropic/claude-opus-4-8`, `anthropic/claude-opus-4-7`, `anthropic/claude-opus-4-6`, `anthropic/claude-sonnet-5`, `anthropic/claude-sonnet-4-6`, `anthropic/claude-haiku-4-5-20251001` are not served on chat completions here: use Anthropic Messages below. A chat completions call for them is refused with HTTP 400. **Base URL:** `https://router.fluxapay.xyz/llm/baidu-ai-cloud/v1` ```bash curl -X POST https://router.fluxapay.xyz/llm/baidu-ai-cloud/v1/chat/completions \ -H "Authorization: Bearer " \ -H "Content-Type: application/json" \ -d '{"model":"","messages":[{"role":"user","content":"..."}]}' ``` Or with the OpenAI SDK: ```ts import OpenAI from "openai" const client = new OpenAI({ baseURL: "https://router.fluxapay.xyz/llm/baidu-ai-cloud/v1", apiKey: process.env.AUTH, // fxa_live_... (human) or Agent VC (agent) }) const res = await client.chat.completions.create({ model: "", messages: [{ role: "user", content: "Hello" }], }) ``` Non-streaming responses include `X-LLM-Cost-Credits` (Units charged for the call) and `X-LLM-Balance` (balance after) headers; streaming responses are metered after the stream closes, without these headers. When unsettled debt crosses the threshold, the **next call returns HTTP 402**. Settle or top up per the wallet guide. ## Anthropic Messages (Claude) `anthropic/claude-opus-5-5`, `anthropic/claude-fable-5`, `anthropic/claude-opus-5`, `anthropic/claude-opus-4-8`, `anthropic/claude-opus-4-7`, `anthropic/claude-opus-4-6`, `anthropic/claude-sonnet-5`, `anthropic/claude-sonnet-4-6`, `anthropic/claude-haiku-4-5-20251001` are also served on Anthropic's native Messages API, with its full feature set (prompt caching with 5-minute and 1-hour writes, thinking, tools). Point any Anthropic SDK or Claude Code at the base URL; the credential goes in `x-api-key` or `Authorization: Bearer`. **Base URL:** `https://router.fluxapay.xyz/llm/baidu-ai-cloud` ```bash curl -X POST https://router.fluxapay.xyz/llm/baidu-ai-cloud/v1/messages \ -H "x-api-key: " \ -H "anthropic-version: 2023-06-01" \ -H "Content-Type: application/json" \ -d '{"model":"","max_tokens":1024,"messages":[{"role":"user","content":"..."}]}' ``` Usage, including `cache_creation_input_tokens` and `cache_read_input_tokens`, is billed at the model's input, cache-write and cache-read rates. ## Wallet Operations for Baidu AI Cloud ```bash # balance (one balance, spends at any merchant) curl -H "Authorization: Bearer " https://router.fluxapay.xyz/llm/wallet/balances # spending history curl -H "Authorization: Bearer " "https://router.fluxapay.xyz/llm/wallet/ledger/baidu-ai-cloud?kind=spend&limit=20" # topup (after explicit user confirmation; vendorSlug optional, bookkeeping only) curl -X POST -H "Authorization: Bearer " -H "Content-Type: application/json" \ -d '{"packageSlug":""}' https://router.fluxapay.xyz/llm/topup/initiate ``` ## Topup Bundles | Bundle | Cost | Units granted | |--------|------|---------------| | Starter (`starter`) | 5 MC ($5) | 500,000 | | Mid (`mid`) | 10 MC ($10) | 1,000,000 | | Pro (`pro`) | 25 MC ($25) | 2,500,000 | **Always confirm a topup with the user before initiating it** — see the wallet guide. ## Units Wallet Guide **Read once — credential setup, all operations, topup protocol, errors:** ```bash curl https://agentmarket.fluxapay.xyz/api/llm/skills.md ``` --- - All merchants: `curl "https://agentmarket.fluxapay.xyz/api/discover?type=model"` - Platform overview: https://agentmarket.fluxapay.xyz/skill.md