Add 'none' reasoning effort for GPT5.1

[machinewrapped]

Added a 'none' reasoning effort to the OpenAI provider settings. Note that this is only supported by GPT5.1.

Conversely, the new gpt-5.1-chat-latest model only supports 'medium' reasoning effort.

Hopefully OpenAI future models or future OpenAI APIs will be less fragile 😓

---

This discussion was created from the release Add 'none' reasoning effort for GPT5.1.

[Neoony]

Funny they released GPT 5.2
And guess what, the gpt-5.2-chat-latest is back to splitting one subtitle "line" into multiple. (perhaps also merging, didnt have that yet)

As if they just reverted back to early 4o
(this got fixed at one point in the 4o chat latest when they updated it, and didnt happen to me since then on any newer models until now with 5.2 chat latest....will still have to try the normal 5.2)
It might even be worse than when it was happening with early 4o / pre 4o, more often. Had it happen almost at every subtitles I tried.

I wonder what did they actually do, since I guess the 5.2 is their "answer" to their "code red" 😄

But at least llm-subtrans seems to work with the gpt-5.2-chat-latest (medium reasoning, whatever that does, that still makes no sense as still nothing says it should reason)

Just to mention :P

[machinewrapped]

Yikes 😆 Maybe time to give Gemini a chance 😉