Repository navigation
GPT-6_luna #928
Replies: 3 comments
|
Thanks for the report, and sorry for the slow reply. You're not doing anything wrong: this is a limitation of Home Generative Agent, not your setup. OpenAI's GPT-6 models only allow tool calls with reasoning through OpenAI's newer Responses API. Home Generative Agent still talks to OpenAI through the older Chat Completions API. It also has no setting today that turns reasoning off for a cloud OpenAI model, which is the other way that error message suggests. So any request that needs a tool (controlling a device, reading a sensor, looking at a camera) fails on a GPT-6 model, whatever you configure. For now: pick a GPT-5 model for the provider's chat model. The fix: GPT-6 support through the Responses API is planned for 3.46.0, the next monthly release, expected in December. I'll post here when it is merged. If you'd like to try it before then, I can offer a pre-release. |
|
@stevenheff GPT-6 support is merged to If you'd like to try it sooner, v3.46.0-beta.1 is out as a pre-release: https://github.com/goruck/home-generative-agent/releases/tag/v3.46.0-beta.1. In HACS, open Home Generative Agent, choose Redownload from the ⋮ menu, turn on Show beta versions, pick After installing, set your Conversation model back to the GPT-6 model you were using. If anything goes wrong, Home Assistant logs from a failing turn would help (Settings → System → Logs, filtered for |
|
Thanks
Have downloaded and that seems to work now.
Will let you know if any issues
…On Fri, 13 Nov 2026, 12:26 am Lindo St. Angel, ***@***.***> wrote:
@stevenheff <https://github.com/stevenheff> GPT-6 support is merged to
main (#993 <#993>)
and ships in *3.46.0*, the next monthly release (around December 9). The
OpenAI provider now uses OpenAI's Responses API for GPT-6, so tool calls
work. GPT-6 also gains Off, Extra-high and Max reasoning settings.
If you'd like to try it sooner, *v3.46.0-beta.1* is out as a pre-release:
https://github.com/goruck/home-generative-agent/releases/tag/v3.46.0-beta.1.
In HACS, open Home Generative Agent, choose *Redownload* from the ⋮ menu,
turn on *Show beta versions*, pick v3.46.0-beta.1, then restart Home
Assistant. When 3.46.0 comes out, HACS will update you to it.
After installing, set your Conversation model back to the GPT-6 model you
were using. If anything goes wrong, Home Assistant logs from a failing turn
would help (Settings → System → Logs, filtered for home_generative_agent).
—
Reply to this email directly, view it on GitHub
<#928?email_source=notifications&email_token=A272WBTQ6ZGAYIYDH4PZTE35YW3CDA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGY2TCNZSUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-19565172>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/A272WBXECUAWXQHQSKGHVSD5YW3CDAVCNFSNUABIKJSXA33TNF2G64TZHM4DIMBTGAZDGMZSHNCGS43DOVZXG2LPNY5TCMJSGQ4DEMBVUF3AE>
.
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS
<https://github.com/notifications/mobile/ios/A272WBQWHAJIZOMMSKHQHID5YW3CDA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGY2TCNZSUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSVGM33PORSXEX3JN5ZQ>
and Android
<https://github.com/notifications/mobile/android/A272WBVFHYSIN3B2NBRXNTL5YW3CDA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOJVGY2TCNZSUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ>.
Download it today!
You are receiving this because you were mentioned.Message ID:
<goruck/home-generative-agent/repo-discussions/928/comments/19565172@
github.com>
|
Uh oh!
There was an error while loading. Please reload this page.
Trying ot use GPT-6 but keep getting an error
I'm sorry, I was unable to respond: Model invocation failed: Error code: 400 - {'error': {'message': "Function tools with reasoning_effort are not supported for gpt-6-luna in /v1/chat/completions. To use function tools, use /v1/responses or set reasoning_effort to 'none'.", 'type': 'invalid_request_error', 'param':…
its fine iof I use a GPT4 model.
I am new to AI so not sure what I am doing wrong?
All reactions