Skip to content

feat: switch to gpt5 models #92#113

Open
bahdan111 wants to merge 52 commits into
developmentfrom
92-switch-to-gpt5-models
Open

feat: switch to gpt5 models #92#113
bahdan111 wants to merge 52 commits into
developmentfrom
92-switch-to-gpt5-models

Conversation

@bahdan111
Copy link
Copy Markdown
Contributor

@bahdan111 bahdan111 commented Jan 20, 2026 •
edited
Loading

Applicable issues

Description of changes

Checklist

By submitting this pull request, I confirm that my contribution is made under the terms of the MIT license.

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Jan 20, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21184340275
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 2, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21583791983

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 2, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21585700433
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 2, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21598795489
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 2, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21600717997
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 3, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21620580989
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 3, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21632109867
Environment URL: review-environment | pipeline

@navalnica navalnica self-requested a review February 3, 2026 13:37
@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 3, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21634222279
Environment URL: review-environment | pipeline

@navalnica navalnica changed the title feat: switch to gpt5 models feat: switch to gpt5 models #92 Feb 3, 2026
@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 3, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21641130555
GitHub actions run: 21641130555
Environment URL: review-environment | pipeline
GitHub actions run: 21641130555
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 4, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21664400101

@navalnica navalnica marked this pull request as ready for review February 5, 2026 11:02
@navalnica navalnica requested a review from ypldan as a code owner February 5, 2026 11:02
@navalnica navalnica marked this pull request as draft February 5, 2026 11:02
@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 5, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21713136521

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 8, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21804121386
Environment URL: review-environment | pipeline

@bahdan111
Copy link
Copy Markdown
Contributor Author

bahdan111 commented Feb 9, 2026 •
edited by statgpt-actions
Loading

/deploy-review

GitHub actions run: 21814096915
Environment URL: review-environment | pipeline

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 18, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26023211582

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277 Azamat0315277 self-assigned this May 18, 2026
@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 19, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26080275829

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 19, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26087929416

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 19, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26095207336

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 20, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26142495794

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 21, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26205456586

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 22, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26272123364

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 25, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26385431216

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 26, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26434134551

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 28, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26555003504

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

@Azamat0315277
Copy link
Copy Markdown
Contributor

Azamat0315277 commented May 29, 2026 •
edited by ai-dial-actions
Loading

/deploy-review

GitHub actions run: 26630064851

Stage Status
deploy-review Success ✅
matrix.application Skipped ➖

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/admin/settings/background_tasks.py
max_concurrent: int = Field(5, description="Maximum number of concurrent background tasks")
task_timeout: float | None = Field(
1800.0, description="Timeout in seconds for a single background task (None to disable)"
3600.0, description="Timeout in seconds for a single background task (None to disable)"
Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why is this needed?

Copy link
Copy Markdown
Contributor Author

@bahdan111 bahdan111 Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

that's fix to timeouts we had during indexing

Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@Fedir-Yatsenko are you okay with this? looks like it affects all background jobs, not only indexing ones

Copy link
Copy Markdown
Collaborator

@Fedir-Yatsenko Fedir-Yatsenko Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yeah, this is okay

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/app/schemas/file_rags/base.py
class BaseRagState(ToolMessageState):
type: ToolTypes = ToolTypes.FILE_RAG
version: RAGVersion
version: RAGVersion = RAGVersion.DIAL
Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why do we set this default?

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/common/schemas/data_query_tool.py
default_factory=lambda: LLMModelConfig(deployment=LLMModelsEnum.GPT_4_1_MINI_2025_04_14),
default_factory=lambda: LLMModelConfig(
deployment=LLMModelsEnum.GPT_5_MINI_2025_08_07,
reasoning_effort=ReasoningEffortEnum.MINIMAL,
Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why not reasoning=none?

Copy link
Copy Markdown
Contributor Author

@bahdan111 bahdan111 Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

there is no none reasoning for mini

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/common/schemas/data_query_tool.py
description="LLM Model used for normalization",
default_factory=lambda: LLMModelConfig(deployment=LLMModelsEnum.GPT_4_1_MINI_2025_04_14),
default_factory=lambda: LLMModelConfig(
deployment=LLMModelsEnum.GPT_5_MINI_2025_08_07,
Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we can try gpt-5.4-mini

Copy link
Copy Markdown
Contributor Author

@bahdan111 bahdan111 Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

for sure

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/common/schemas/data_query_tool.py
Comment on lines +188 to +191
deployment=LLMModelsEnum.GPT_5_MINI_2025_08_07,
reasoning_effort=ReasoningEffortEnum.MINIMAL,
verbosity=VerbosityEnum.LOW,
temperature=1,
Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same. can try gpt-5.4-mini and none reasoning

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/common/settings/langchain.py
Comment on lines +46 to +55
default_reasoning_effort: ReasoningEffortEnum | None = Field(
default=ReasoningEffortEnum.NONE,
description="Default reasoning effort for GPT-5 models (none/minimal/low/medium/high/xhigh)",
)

default_verbosity: VerbosityEnum | None = Field(
default=VerbosityEnum.LOW,
description="Default verbosity for GPT-5 models (low/medium/high). None means use model default.",
)

Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

are those ever used?

navalnica
navalnica reviewed Jun 3, 2026
View reviewed changes
Comment thread statgpt/common/utils/callbacks.py
Copy link
Copy Markdown
Contributor

@navalnica navalnica Jun 3, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

all changes to this whole file look stale. need to remove them

@navalnica navalnica marked this pull request as ready for review June 3, 2026 08:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@Fedir-Yatsenko Fedir-Yatsenko Fedir-Yatsenko left review comments

@navalnica navalnica navalnica requested changes

@ypldan ypldan Awaiting requested review from ypldan ypldan is a code owner

Requested changes must be addressed to merge this pull request.

Assignees

@bahdan111 bahdan111

@Azamat0315277 Azamat0315277

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

Core: Switch to GPT-5 series models

7 participants

@bahdan111 @ypldan @MykhailoRyzhman @Azamat0315277 @navalnica @Fedir-Yatsenko @github-advanced-security