LiteLLM中supports_parallel_function_calling返回值与文档不符问题

我在Python项目中使用LiteLLM测试不同模型对*并行函数调用（parallel function calling）*的支持情况，编写了最小复现代码：

from litellm import completion
from litellm.utils import supports_function_calling, supports_parallel_function_calling

models = [
    "gpt-5",
    "gpt-5-mini",
    "gpt-4-turbo-preview",
    "gpt-4o",
    "gpt-3.5-turbo-1106"
]

for model in models:
    print(f"Model: {model}")
    print("  Supports function calling:", supports_function_calling(model))
    print("  Supports parallel function calling:", supports_parallel_function_calling(model))
    print()

运行后输出：

Model: gpt-5
  Supports function calling: True
  Supports parallel function calling: False

Model: gpt-5-mini
  Supports function calling: True
  Supports parallel function calling: False

Model: gpt-4-turbo-preview
  Supports function calling: True
  Supports parallel function calling: False

Model: gpt-4o
  Supports function calling: True
  Supports parallel function calling: False

Model: gpt-3.5-turbo-1106
  Supports function calling: True
  Supports parallel function calling: False

根据LiteLLM官方文档，以下断言应当成立：

assert litellm.supports_parallel_function_calling(model="gpt-4-turbo-preview") == True

但实际测试中，supports_parallel_function_calling("gpt-4-turbo-preview")始终返回False，与预期不符。当前使用的LiteLLM版本为1.76.0，即使在模型名称中指定提供商（如openai/gpt-4-turbo-preview），问题仍未解决，官方文档提及的gpt-3.5-turbo-1106模型也存在同样问题。

• 升级LiteLLM版本
  LiteLLM的模型支持列表可能在后续版本中更新，尝试升级到最新稳定版：

  pip install --upgrade litellm
  

  升级后重新运行测试代码，确认函数返回值是否符合预期。

• 手动验证模型实际能力
  绕过supports_parallel_function_calling工具函数，直接调用模型发起并行函数调用请求，验证模型本身是否支持该功能。示例代码：

  from litellm import completion
  
  tools = [
      {
          "type": "function",
          "function": {
              "name": "get_current_weather",
              "description": "Get the current weather in a given location",
              "parameters": {
                  "type": "object",
                  "properties": {
                      "location": {
                          "type": "string",
                          "description": "The city and state, e.g. San Francisco, CA"
                      },
                      "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
                  },
                  "required": ["location"]
              }
          }
      },
      {
          "type": "function",
          "function": {
              "name": "get_news",
              "description": "Get the latest news for a given topic",
              "parameters": {
                  "type": "object",
                  "properties": {
                      "topic": {
                          "type": "string",
                          "description": "The news topic, e.g. technology"
                      }
                  },
                  "required": ["topic"]
              }
          }
      }
  ]
  
  response = completion(
      model="gpt-4-turbo-preview",
      messages=[{"role": "user", "content": "告诉我北京的天气和最新的科技新闻"}],
      tools=tools,
      tool_choice="auto"
  )
  
  # 检查是否返回多个工具调用
  if response.choices[0].message.tool_calls and len(response.choices[0].message.tool_calls) > 1:
      print("模型支持并行函数调用")
  else:
      print("模型不支持并行函数调用")
  

  如果实际调用成功返回多个工具调用，说明是supports_parallel_function_calling函数的判断逻辑存在问题。

• 检查LiteLLM源码配置
  查看LiteLLM源码中的模型支持配置，确认目标模型是否被标记为支持并行函数调用。可以在LiteLLM的model_info或类似配置文件中查找对应模型的属性，若未正确配置，可提交Issue反馈。

内容的提问来源于stack exchange，提问作者EnzoDeg40