One endpoint, every model
A single https://api.zaimod.com/v1 — swap the model name to compare vendors without rewriting code.
ChatGPT, Gemini, Claude, DeepSeek, Doubao, Qwen, Kimi and GLM behind a single endpoint. Change one base_url line, leave the rest of your code alone.
from openai import OpenAI
client = OpenAI(
api_key="sk-your-zaimod-key",
base_url="https://api.zaimod.com/v1",
)
resp = client.chat.completions.create(
model="gpt-4o-mini", # swap for claude-* / gemini-* / deepseek-*
messages=[{"role": "user", "content": "Explain API gateways in three sentences"}],
)
print(resp.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.ZAIMOD_API_KEY,
baseURL: "https://api.zaimod.com/v1",
});
const res = await client.chat.completions.create({
model: "claude-sonnet-5",
messages: [{ role: "user", content: "Summarise this changelog in 5 bullets." }],
});
console.log(res.choices[0].message.content);curl https://api.zaimod.com/v1/chat/completions \
-H "Authorization: Bearer $ZAIMOD_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gemini-3.1-flash-lite-preview",
"messages": [{"role": "user", "content": "Hello"}],
"stream": true
}'
Keep your tooling. Turn multiple platforms, keys and invoices into one.
A single https://api.zaimod.com/v1 — swap the model name to compare vendors without rewriting code.
Beyond OpenAI format, Claude /v1/messages and Gemini /v1beta/models work natively, so official SDKs stay first-class.
Token usage and charges are itemised per key and per model in the console, priced from the same table you see here.
The default group covers every model; official and premium groups add stability for specific vendors, with the multiplier published.
Any SDK, plugin or client that speaks OpenAI works out of the box.
Switch groups with the same key. Multiplier applies to the base price and is visible per call.
纯Claude官转 · 比官方便宜40%
Gemini 优质 · 支持视频分析 · 比官方便宜80%
纯OpenAI特价优质 · 比官方便宜45%
默认分组首选 · 支持所有模型便宜80%
Synced from the platform pricing API. USD per 1M tokens; fixed-price models bill per call.
| Model | Vendor | Input | Output | Tags |
|---|---|---|---|---|
| gpt-4o-mini | OpenAI | $0.156 / $0.624 | 对话 | |
| gpt-4.1 | OpenAI | $2 / $4 | 编程对话 | |
| gpt-5.6-terra | OpenAI | $3 / $15 | 编程对话多模态 | |
| gpt-6-astra | OpenAI | $10 / $50 | 对话推理智能体 | |
| claude-sonnet-5 | Anthropic Claude | $4 / $20 | 对话Agent智能体 | |
| claude-opus-4-7 | Anthropic Claude | $8 / $40 | 对话多模态智能体 | |
| gemini-3.1-flash-lite-preview | Google Gemini | $0.25 / $1.5 | 对话推理 | |
| gemini-3.1-pro-preview | Google Gemini | $4 / $24 | 对话 | |
| deepseek-v4-flash | DeepSeek | $3 / $9 | 对话 | |
| doubao-seed-2-0-pro-260215 | 火山豆包 | $5 / $24 | 对话智能体 | |
| qwen3.5-plus | 通义千问 | $0.8 / $5 | 推理对话 | |
| glm-5.2 | 智谱 GLM | $8 / $28 | 编程智能体 |
data synced 2026-11-03 · Prices are the platform list price at the default group (multiplier 1.0). Official and premium groups use a 1.5 multiplier; the console is authoritative.
One key from code editor to production.
Drop the custom Base URL into Cursor, Cline, Roo Code or Continue and move between Claude and GPT models at will.
Power WeChat, Telegram or Discord bots: cheap models for routine answers, premium models when it gets hard.
Translate, summarise and extract in parallel. Token-based pricing beats subscriptions for spiky workloads.
Price comparisons, integration guides and field notes, updated weekly.
deepseek-v4-flash-free is $0 per call with daily rate limits. Five production uses: intent classification, tagging, JSON extraction, summaries, batch jobs.
LLM streaming is just an SSE long HTTP connection: event-stream format, stream:true chunks, curl and frontend debugging, and classic proxy-buffering pitfalls.
Gateway shutdowns are never sudden. Balance rules, response fingerprints, update cadence, support silence and payment rails — five measurable warning signs.
From your first call to production debugging.
From signup to your first API call — create a key, swap the base_url, pick a model, with minimal working examples for the OpenAI, Claude and Gemini protocols.
DOCCall all 175 models through the OpenAI-compatible endpoint — streaming, tool calling, multimodal image input, structured JSON output and embeddings, with copy-paste examples.
DOCCall the full Claude lineup over the native Anthropic protocol — message structure, the system parameter, extended thinking, native tool use and multi-turn conversations.
DOCCall the Gemini lineup over the native Google protocol — generateContent structure, multimodal parts, system instructions, thinking budgets and streaming.
Yes. Chat completions follow the OpenAI /v1/chat/completions spec, so official SDKs only need a new base_url and api_key. Claude /v1/messages and Gemini /v1beta/models are also exposed natively.
Per token: model base price × group multiplier. Image, video and 3D models bill per call with a flat unit price. Every call is itemised in the console.
Top-up credits have no expiry date. The console billing page is the source of truth.
SSE streaming, function calling / tools, multimodal image input and structured JSON output are supported. Exact support depends on the model.
Multi-channel routing with automatic failover; official groups route to official channels. Live status and latency are available in the console.
Contact us via live chat or email with your usage and billing entity, and we will propose a settlement method.
Trial credits on signup. No overseas card needed.