AIHubMixAIHubMix
智能路由model = "auto"

并不是每一次请求，都需要最强最贵的模型

智能路由为每个请求实时选出最佳模型，在不影响任务质量的前提下大幅降低成本；模型缓存支持，同一会话自动保持同一模型，避免缓存丢失。

无需额外改造：Claude Code、Codex、OpenCode 等主流 Agent 的会话标识会被自动识别。需要精确控制会话范围时，可在请求中添加：X-Aihubmix-Session-Id: your-session-id

评分覆盖 5 大类、23 个子维度，来自内部测评与外部公开榜单的综合归一化。

接口文档
chat.py
import openai

client = openai.OpenAI(
    api_key="<AIHUBMIX_API_KEY>",
    base_url="https://aihubmix.com/v1"
)

response = client.chat.completions.create(
    model="auto",  # 不带后缀 = auto:cost_optimized（默认）
    messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)
路由策略
路由策略

四大路由策略对应的模型范围

文本 › 总体

按「成本优先」策略加权排序
模型
18
A
claude-opus-4-6-thinkAnthropic
86.7
$2.5 / $51,673 ms
21
A
claude-opus-4-7-thinkAnthropic
86.3
$2.5 / $53,036 ms
14
A
claude-opus-4-6Anthropic
85.4
$2.5 / $5822 ms
19
A
claude-opus-4-7Anthropic
84.0
$2.5 / $5719 ms
26
G
gemini-3.1-pro-previewGoogle
81.7
$1 / $66,816 ms
20
G
gemini-3.5-flashGoogle
79.9
$0.75 / $64,328 ms
28
O
gpt-5.4-highOpenAI
79.4
$1.25 / $6—
24
A
claude-opus-4-8Anthropic
79.0
$2.5 / $51,591 ms
29
O
gpt-5.5OpenAI
78.5
$2.5 / $64,766 ms
8
X
grok-4.20-beta-0309-reasoningxAI
78.5
$1 / $31,685 ms
11
A
claude-sonnet-4-6Anthropic
77.8
$1.5 / $5995 ms
12
A
claude-sonnet-4-6-thinkAnthropic
77.3
$1.5 / $51,762 ms
9
Z
glm-5.2Z.ai
76.8
$0.563 / $3.52,770 ms
1
M
xiaomi-mimo-v2.5-pro当前首选Xiaomi
76.1
$0.24 / $21,673 ms
10
Q
qwen3.7-plusQwen
75.0
$0.141 / $44,873 ms
13
Q
qwen3.6-max-previewQwen
74.4
$0.634 / $6651 ms
4
X
grok-4-1-fast-non-reasoningxAI
74.2
$0.1 / $2.51,188 ms
2
D
deepseek-v4-proDeepSeek
73.5
$0.232 / $22,037 ms
17
O
gpt-5.4-miniOpenAI
71.0
$0.375 / $61,354 ms
6
X
grok-4.3xAI
69.7
$0.625 / $21,224 ms
22
Q
qwen3.6-plusQwen
69.5
$0.141 / $64,120 ms
16
Q
qwen3.5-397b-a17bQwen
69.4
$0.082 / $62,106 ms
3
D
deepseek-v4-flashDeepSeek
67.3
$0.077 / $2879 ms
5
M
xiaomi-mimo-v2.5Xiaomi
66.6
$0.078 / $21,673 ms
15
G
gemini-3.1-flash-liteGoogle
66.4
$0.125 / $6733 ms
7
X
grok-4-1-fast-reasoningxAI
66.0
$0.1 / $2.52,003 ms
30
Q
qwen3.5-122b-a10bQwen
62.0
$0.056 / $8961 ms
31
Q
qwen3.5-27bQwen
59.7
$0.042 / $82,085 ms
25
O
gpt-5.4-nanoOpenAI
57.8
$0.1 / $6.25702 ms
33
Q
qwen3.5-flashQwen
56.5
$0.014 / $105,775 ms
32
Q
qwen3.5-35b-a3bQwen
55.9
$0.028 / $83,310 ms
23
O
gpt-oss-120bOpenAI
43.6
$0.09 / $5366 ms
27
O
gpt-oss-20bOpenAI
33.5
$0.055 / $5120 ms
数据来源：内部测评与外部公开榜单，综合归一化。
推理时代

© 2023 - 2026 AIHubMix, LLC

产品

公司

开发者

联系我们