A request that sends temperature can come back with the sentence below. OpenAI's Using GPT-6 guide says to remove temperature when reasoning effort is not none.
Unsupported parameter: 'temperature' is not supported with this model.Quoted from GitHub issue neondatabase/neon-pkgs#937, checked October 24, 2026.
The Using GPT-6 guide says: when reasoning effort is not none, remove temperature, top_p, and top_logprobs. For Chat Completions, also remove logprobs. For Responses, remove message.output_text.logprobs from include. GPT-6 Astra and GPT-6.1 Sol do not support reasoning effort none. GPT-6 Sol and GPT-6 Luna do. https://github.com/neondatabase/neon-pkgs/issues/937 records the temperature sentence for gpt-6-astra, gpt-6-sol, gpt-6-luna, and gpt-6-1-sol when temperature is set and no reasoning effort is set. The same issue records this sentence for top_p on gpt-6-astra: Unsupported parameter: 'top_p' is not supported with this model. Its table says gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, gpt-6-sol, and gpt-6-luna accept temperature and top_p only with effort none. https://github.com/maxswaine/game-repository/issues/110 says gpt-4.1-nano shut down on 2026-10-23, that calls return 404 model_not_found, and that temperature is unsupported on successor models. That issue does not print the temperature sentence. SunsetDiff's gpt-4.1-nano record lists shutdown date 2026-10-23 and replacement gpt-5.6-luna. gpt-6-astra and gpt-6.1-sol are named in the guide and in the neon issue. SunsetDiff has no model file for those two ids, so they are not linked below. Source: Using GPT-6, checked October 24, 2026.
| Model id | What the source says |
|---|---|
gpt-6-sol | The neon issue reproduced the temperature sentence on this id. The Using GPT-6 guide says GPT-6 Sol supports reasoning effort none, and to remove temperature when effort is not none. |
gpt-6-luna | The neon issue reproduced the temperature sentence on this id. The Using GPT-6 guide says GPT-6 Luna supports reasoning effort none. |
gpt-5.6-sol | The neon issue's table says this id accepts temperature and top_p only with effort none. The quoted sentence in that issue was reproduced for the gpt-6 ids, not for this id. |
gpt-5.6-luna | SunsetDiff lists this id as the replacement for gpt-4.1-nano, which shut down October 23, 2026. The neon issue's table says this id accepts temperature and top_p only with effort none. The quoted sentence was not reproduced for this id in that issue. |
gpt-5.6-terra | The neon issue's table says this id accepts temperature and top_p only with effort none. The quoted sentence was not reproduced for this id in that issue. |
Omit temperature when reasoning effort is not none. gpt-6-sol and gpt-6-luna can keep temperature if reasoning_effort is none. GPT-6 Astra and GPT-6.1 Sol do not support none, so the guide says to remove temperature on those models. The sample uses gpt-6-sol, which is in this dataset.
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gpt-6-sol",
messages=[{"role": "user", "content": "Hello"}],
temperature=0.2,
)
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gpt-6-sol",
messages=[{"role": "user", "content": "Hello"}],
)
response = client.chat.completions.create(
model="gpt-6-sol",
messages=[{"role": "user", "content": "Hello"}],
reasoning_effort="none",
temperature=0.2,
)
Source: Using GPT-6, checked October 24, 2026.
Do not send temperature. additional_drop_params removes named fields on that call. This page does not claim LiteLLM already lists temperature as unsupported for gpt-6-sol.
import litellm
response = litellm.completion(
model="openai/gpt-6-sol",
messages=[{"role": "user", "content": "Hello"}],
additional_drop_params=["temperature", "top_p"],
)
Source: LiteLLM drop unsupported params, checked October 24, 2026.
The neon issue called generateText with temperature 0.2 and recorded the sentence above for the gpt-6 ids. Leave temperature unset. The generateText reference says temperature is passed through to the provider. That reference does not print this error sentence.
import { openai } from "@ai-sdk/openai";
import { generateText } from "ai";
const { text } = await generateText({
model: openai("gpt-6-sol"),
prompt: "Hello",
});
Source: AI SDK generateText, checked October 24, 2026.
neondatabase/neon-pkgs#937 records it for gpt-6-astra, gpt-6-sol, gpt-6-luna, and gpt-6-1-sol. gpt-6-astra and gpt-6.1-sol have no SunsetDiff model file. The same issue records a matching top_p sentence for gpt-6-astra.
The Using GPT-6 guide says GPT-6 Sol and GPT-6 Luna support reasoning effort none, and to remove temperature when effort is not none. GPT-6 Astra and GPT-6.1 Sol do not support none.
maxswaine/game-repository#110 records 404 model_not_found for gpt-4.1-nano on October 23, 2026, and says temperature is unsupported on the successor. That issue does not print the temperature sentence. The shutdown sentence is on The model has been deprecated.