自定义供应商 中转站 提示不支持 xhigh

[Monster-DP]

不是哟 cliproxyapi 直接使用中转站是没有问题的

{"error":{"message":"level "xhigh" not supported, valid levels: low, medium, high","type":"invalid_request_error"}}

[luispater]

走的是openai chat completions协议么？参考下面的范例：

openai-compatibility:
 - name: "openrouter"
   base-url: "https://example.com/v1"
   api-key-entries:
     - api-key: "sk-or-v1-...b780"
   models: # The models supported by the provider.
     - name: "gpt-5.4"
       alias: "gpt-5.4"
       thinking:
         levels: ["low", "medium", "high", "xhigh"] # 这里添加xhigh就可以了

[Monster-DP]

走的是openai chat completions协议么？参考下面的范例：

openai-compatibility:
 - name: "openrouter"
   base-url: "https://example.com/v1"
   api-key-entries:
     - api-key: "sk-or-v1-...b780"
   models: # The models supported by the provider.
     - name: "gpt-5.4"
       alias: "gpt-5.4"
       thinking:
         levels: ["low", "medium", "high", "xhigh"] # 这里添加xhigh就可以了

我是用的 codex 直接配置的 ccswitch
再codex中使用 5.4 xhigh 就会出现这个 {"error":{"message":"level "xhigh" not supported, valid levels: low, medium, high","type":"invalid_request_error"}}
应该是用的 OpenAI Response 这个请求格式吧

[luispater]

如果走的是responses协议，你需要将密码之类的信息填写在codex-api-key段下，如果填在openai-compatibility下，就是走openai chat completions的

[Monster-DP]

codex-api-key

非常感谢你！！❤️❤️❤️❤️

[DeepCoach]

   models: # The models supported by the provider.
     - name: "gpt-5.4"
       alias: "gpt-5.4"
       thinking:
         levels: ["low", "medium", "high", "xhigh"] # 这里添加xhigh就可以了

像这个openai-compatibility如果设置了多个模型（如gpt5.4/5.5）,是不是每个想要支持xhigh模型的alias下都得填一遍：

thinking:
  levels: ["low", "medium", "high", "xhigh"]