Function tools with reasoning_effort are not supported for gpt-6-luna

Raw SDK

For function tools, call the Responses API. To stay on Chat Completions, set reasoning_effort to none. The model page says that is the only Chat Completions function-calling mode for gpt-6-luna. The Responses tool object matches OpenAI's function calling guide: https://developers.openai.com/api/docs/guides/function-calling

from openai import OpenAI
client = OpenAI()
response = client.responses.create(
    model="gpt-6-luna",
    input=[{"role": "user", "content": "Hello"}],
    tools=[{
        "type": "function",
        "name": "get_weather",
        "parameters": {
            "type": "object",
            "properties": {"city": {"type": "string"}},
            "required": ["city"],
        },
    }],
)

from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
    model="gpt-6-luna",
    messages=[{"role": "user", "content": "Hello"}],
    reasoning_effort="none",
    tools=[{
        "type": "function",
        "function": {
            "name": "get_weather",
            "parameters": {
                "type": "object",
                "properties": {"city": {"type": "string"}},
                "required": ["city"],
            },
        },
    }],
)

Source: GPT-6 Luna model page, checked October 20, 2026.

LiteLLM

LiteLLM documents model openai/responses/gpt-6-luna for the Responses API. Use that when the call includes function tools. On a Chat Completions model string, set reasoning_effort to none, which is the only function-calling mode the OpenAI model page allows on that endpoint.

import litellm
response = litellm.completion(
    model="openai/responses/gpt-6-luna",
    messages=[{"role": "user", "content": "Hello"}],
    tools=[{
        "type": "function",
        "function": {
            "name": "get_weather",
            "parameters": {
                "type": "object",
                "properties": {"city": {"type": "string"}},
                "required": ["city"],
            },
        },
    }],
)

import litellm
response = litellm.completion(
    model="openai/gpt-6-luna",
    messages=[{"role": "user", "content": "Hello"}],
    reasoning_effort="none",
    tools=[{
        "type": "function",
        "function": {
            "name": "get_weather",
            "parameters": {
                "type": "object",
                "properties": {"city": {"type": "string"}},
                "required": ["city"],
            },
        },
    }],
)

Source: LiteLLM OpenAI provider, checked October 20, 2026.

Vercel AI SDK

openai('gpt-6-luna') uses the Responses API. The AI SDK guide says to use the Responses API for GPT-6 tool calling. openai.chat() stays on Chat Completions, so set providerOptions.openai.reasoningEffort to none when that call includes tools.

import { openai } from "@ai-sdk/openai";
import { generateText, tool } from "ai";
import { z } from "zod";

const result = await generateText({
  model: openai("gpt-6-luna"),
  prompt: "What is the weather in Paris?",
  tools: {
    get_weather: tool({
      description: "Get the weather for a city.",
      inputSchema: z.object({ city: z.string() }),
    }),
  },
});

import { openai, type OpenAILanguageModelChatOptions } from "@ai-sdk/openai";
import { generateText, tool } from "ai";
import { z } from "zod";

const result = await generateText({
  model: openai.chat("gpt-6-luna"),
  prompt: "What is the weather in Paris?",
  tools: {
    get_weather: tool({
      description: "Get the weather for a city.",
      inputSchema: z.object({ city: z.string() }),
    }),
  },
  providerOptions: {
    openai: {
      reasoningEffort: "none",
    } satisfies OpenAILanguageModelChatOptions,
  },
});

Source: AI SDK OpenAI guide, checked October 20, 2026.

Questions