对话

POST /v1/chat/completions

在文本模型上创建一次对话补全。请求和响应兼容 OpenAI；同步返回——回复就在响应里（或流式返回，见下）。积分按预估 prompt 加上 min(max_tokens, default_max_tokens) 个输出 token 预留，调用结束时按实际用量结算；没有产生 token 则不扣费。

请求头

请求头	必填	说明
Authorization	是	Bearer <API 密钥>
Content-Type	是	application/json
Idempotency-Key	否	这里的 Idempotency-Key 是可选的。重复使用会返回 `409 idempotency_key_reused`.

请求体

字段	类型	说明
model	string	必填。来自 `GET /v1/models` 的文本模型 id，例如 `"qwen3.8-max"`。
messages	array	必填。`{role: "system" \| "user" \| "assistant", content: string \| parts[]}`。内容片段（`text`、`image_url`）与 OpenAI API 一致；`image_url` 仅在 `vision: true` 的模型上可用。
max_tokens	integer	回复上限。不超过模型的 `max_output_tokens`；默认为 `default_max_tokens`。推理 token 也计入其中。
temperature	number	0–2。默认取模型的 default_params。
top_p	number	0–1。
stop	string[]	最多 4 个停止序列。
stream	boolean	默认 `false`。设为 `true` → `text/event-stream`，见「流式输出」。
enable_thinking	boolean	扩展参数。作答前先推理，结果以 `reasoning_content` 返回。仅在 `thinking: true` 的模型上可用。
web_search	boolean	扩展参数。允许模型查阅网络。仅在 `web_search: true` 的模型上可用。
reasoning_effort	string	`none` / `minimal` 关闭思维链，`low` … `xhigh` 开启。仅在 `thinking: true` 的模型上可用。
tools	array	OpenAI 函数定义。仅在 `tools: true` 的模型上可用；见工具调用。
tool_choice	string \| object	`auto` · `none` · `required` · `{type: "function", function: {name}}`。
response_format	object	`{type: "json_object"}` 或 `{type: "json_schema", json_schema: {name, schema}}`；见结构化输出。

不支持——会以 400 拒绝，unsupported_parameter，绝不静默忽略： n > 1, functions (legacy), logprobs.

curl -s -X POST https://api.genmux.tech/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen3.8-max","messages":[{"role":"user","content":"Explain idempotency keys in two sentences."}],"max_tokens":200}'

响应 `200 OK`

{
  "id": "0d8b6a3e-2f1c-4a5b-9c1d-7e2f3a4b5c6d",
  "object": "chat.completion",
  "created": 1756411279,
  "model": "qwen3.8-max",
  "choices": [
    {"index": 0, "message": {"role": "assistant", "content": "An idempotency key is …"}, "finish_reason": "stop"}
  ],
  "usage": {"prompt_tokens": 31, "completion_tokens": 58, "total_tokens": 89, "cost": 0.004, "cost_source": "vendor"}
}

usage.cost 是本次扣除的积分（小数，最多 3 位）；cost_source 在上游报告 token 数时为 "vendor"，"estimated" 表示是本地统计的。

限制

字段	限制
max_tokens	不超过模型的 max_output_tokens；默认取 default_max_tokens。若 prompt 长到没有空间生成回复，会返回 400 context_length_exceeded。
stop	最多 4 个字符串。
temperature	0–2。
messages	上下文窗口内任意数量的 system、user 和 assistant 轮次。image_url 部分仅在 vision: true 的模型上可用。

流式输出

设置 "stream": true 后，响应是 text/event-stream。每一行 data: 携带一个 chat.completion.chunk；最后一个分块的 choices 数组为空并带上 usage 块（含本次扣除的积分），流以此结束：data: [DONE].

curl -sN -X POST https://api.genmux.tech/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen3.8-max","messages":[{"role":"user","content":"Explain idempotency keys in two sentences."}],"stream":true}'

Response · text/event-stream

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[{"index":0,"delta":{"role":"assistant","content":"An"},"finish_reason":null}]}

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[{"index":0,"delta":{"content":" idempotency"},"finish_reason":null}]}

…

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[],"usage":{"prompt_tokens":31,"completion_tokens":58,"total_tokens":89,"cost":0.004,"cost_source":"estimated"}}

data: [DONE]

若上游在流中途失败，会发送一帧错误代替 usage 分块，流仍以 [DONE] 结束；预留按实际生成的内容结算，GET /v1/generations/{id} 会显示最终费用。

Mid-stream failure

data: {"error": {"code": "upstream_error", "message": "the model provider closed the connection"}}

data: [DONE]

关闭连接即停止生成。中断前已生成的 token 会计费，其余不计。

OpenAI SDK

该接口使用 OpenAI 的 chat completions 方言，因此官方 SDK 无需改动即可使用：把 base_url 指向 https://api.genmux.tech/v1，并把你的密钥作为 api_key 传入。模型 id 用我们的（见 GET /v1/models），不是厂商的。

curl -s -X POST https://api.genmux.tech/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen3.8-max","messages":[{"role":"user","content":"Explain idempotency keys in two sentences."}],"max_tokens":200}'

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.genmux.tech/v1", api_key=os.environ["API_KEY"])

completion = client.chat.completions.create(
    model="qwen3.8-max",
    messages=[{"role": "user", "content": "Explain idempotency keys in two sentences."}],
    max_tokens=200,
)
print(completion.choices[0].message.content)
print(completion.usage.model_extra["cost"], "credits")

stream = client.chat.completions.create(model="qwen3.8-max", messages=[{"role": "user", "content": "Hi"}], stream=True)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")
    if chunk.usage:
        print("\n", chunk.usage.total_tokens, "tokens,", chunk.usage.model_extra["cost"], "credits")

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.genmux.tech/v1", apiKey: process.env.API_KEY });

const completion = await client.chat.completions.create({
  model: "qwen3.8-max",
  messages: [{ role: "user", content: "Explain idempotency keys in two sentences." }],
  max_tokens: 200,
});
console.log(completion.choices[0].message.content, completion.usage.cost);

const stream = await client.chat.completions.create({ model: "qwen3.8-max", messages: [{ role: "user", content: "Hi" }], stream: true });
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
  if (chunk.usage) console.log(`\n${chunk.usage.total_tokens} tokens, ${chunk.usage.cost} credits`);
}

支持：model、messages（system/user/assistant；视觉模型上的 image_url 部分）、max_tokens、temperature、top_p、stop、stream，以及作为额外请求体字段传入的扩展 enable_thinking 和 web_search。

同样支持：在 tools: true 的模型上使用 tools / tool_choice（见「工具调用」）、response_format 的 json_object / json_schema（见「结构化输出」）、reasoning_effort。会以 400 unsupported_parameter 拒绝：n > 1、旧版 functions 字段、logprobs。接受但忽略：stream_options、parallel_tool_calls、seed、user、metadata、store、logit_bias 以及 message 的 name。

从补全结果（同步）或最后一个分块（流式）读取 usage：cost 就是本次扣除的积分。

工具调用

在目录条目标注 tools: true 的模型上支持 OpenAI 函数调用。发送 tools + tool_choice（auto · none · required · {type: "function", function: {name}}）；回复会带上 message.tool_calls，finish_reason 为 tool_calls；用带有对应 tool_call_id 的 role: tool 消息把结果传回。函数调用由本 API 模拟实现——工具被渲染进 system prompt，回复再解析回来——因此线上格式与 OpenAI 完全一致，而模型本身从未见过原生的 tools 字段。

当前 tools: true 的模型：doubao-seed-2-pro, doubao-seed-2-code, kimi-k3, claude-opus-5-5, claude-fable-5-1, gpt-6-astra, gpt-5.6-terra, qwen3.8-max. 向其他模型发送 tools 会返回 400 tools_unsupported。

Python · OpenAI SDK

from openai import OpenAI
client = OpenAI(base_url="https://api.genmux.tech/v1", api_key="nb_live_...")
tools = [{"type": "function", "function": {"name": "get_weather", "description": "Weather by city",
          "parameters": {"type": "object", "properties": {"city": {"type": "string"}}, "required": ["city"]}}}]
r = client.chat.completions.create(model="qwen3.8-max",
        messages=[{"role": "user", "content": "Weather in Paris?"}], tools=tools)
call = r.choices[0].message.tool_calls[0]        # call.function.name == "get_weather"
# run your function, then send the result back:
r2 = client.chat.completions.create(model="qwen3.8-max", tools=tools, messages=[
    {"role": "user", "content": "Weather in Paris?"},
    r.choices[0].message,
    {"role": "tool", "tool_call_id": call.id, "content": "18°C, cloudy"},
])
print(r2.choices[0].message.content)

Response · first call

{
  "id": "0d8b6a3e-…",
  "object": "chat.completion",
  "model": "qwen3.8-max",
  "choices": [
    {"index": 0, "finish_reason": "tool_calls", "message": {"role": "assistant", "content": null,
      "tool_calls": [{"id": "call_1", "type": "function", "function": {"name": "get_weather", "arguments": "{\"city\": \"Paris\"}"}}]}}
  ],
  "usage": {"prompt_tokens": 96, "completion_tokens": 21, "total_tokens": 117, "cost": 0.003, "cost_source": "vendor"}
}

设置 stream: true 后，调用会以 delta.tool_calls 分块到达——先是 index、id 和名称，然后是分片的参数——随后是一个 finish_reason 为 tool_calls 的分块和照常的 usage 分块。

Streaming · text/event-stream

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[{"index":0,"delta":{"role":"assistant","tool_calls":[{"index":0,"id":"call_1","type":"function","function":{"name":"get_weather","arguments":""}}]},"finish_reason":null}]}

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[{"index":0,"delta":{"tool_calls":[{"index":0,"function":{"arguments":"{\"city\": \"Paris\"}"}}]},"finish_reason":null}]}

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[{"index":0,"delta":{},"finish_reason":"tool_calls"}]}

data: {"id":"0d8b6a3e-…","object":"chat.completion.chunk","model":"qwen3.8-max","choices":[],"usage":{"prompt_tokens":96,"completion_tokens":21,"total_tokens":117,"cost":0.003,"cost_source":"estimated"}}

data: [DONE]

限制

字段	限制
tools	最多 128 个定义，总量不超过 64 KiB。每个定义都会消耗 prompt token：约 40 个，外加它的 JSON schema。
tool_choice	auto（默认）、none、required 或指定函数。required 和指定函数是模型通常会遵守的指令；小模型偶尔会改用自然语言作答。
并行调用	模型通常每轮只发一个工具调用；parallel_tool_calls 会被接受但忽略。智能体是循环工作的，多步任务依然能完成。
strict	接受但忽略——上游没有受约束解码。
response_format	不能与 tools 在同一个请求中同时使用。
max_tokens	超过 max_output_tokens 的值会被收敛，而不是拒绝。工具模型的默认回复预算是 8192 token。

OpenCode、Aider、Cline、Roo Code 和 Continue 的实测配置，以及测得的通过率，见编程智能体页面 →

结构化输出

传入 response_format，取值 {"type": "json_object"} 或 {"type": "json_schema", "json_schema": {"name", "schema"}}（{name, schema}，JSON Schema 2020-12），模型会被要求只用 JSON 作答。非流式回复会被校验：失败时 API 会把校验错误反馈给模型并自动重试一次；第二次仍失败则返回 502 json_invalid，两次尝试都计费。流式回复不做校验。response_format 与 tools 不能同时使用。

curl -s -X POST https://api.genmux.tech/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen3.8-max",
       "messages":[{"role":"user","content":"Extract the city and the date: meet me in Paris on 3 March"}],
       "response_format":{"type":"json_schema","json_schema":{"name":"event",
         "schema":{"type":"object","properties":{"city":{"type":"string"},"date":{"type":"string"}},"required":["city","date"],"additionalProperties":false}}}}'

响应 · content 就是那份 JSON 文档

{"id": "…", "object": "chat.completion", "model": "qwen3.8-max", "choices": [{"index": 0, "message": {"role": "assistant", "content": "{\"city\": \"Paris\", \"date\": \"2026-03-03\"}"}, "finish_reason": "stop"}], "usage": {…}}