langchain-ai/docs
1
0
Fork 0
mirror of https://github.com/langchain-ai/docs.git synced 2026-07-22 04:15:25 -04:00
Code Issues 490 Packages Projects Releases Wiki Activity

[GH-ISSUE #3730] [deepagents]: Document OpenAI Responses API default on models page #2753

New Issue
Open
opened 2026-06-05 17:26:37 -04:00 by yindo · 0 comments
yindo commented 2026-06-05 17:26:37 -04:00
Owner
Copy Link
Copy Source

Originally created by @AdemBoukhris457 on GitHub (Apr 25, 2026).
Original GitHub issue: https://github.com/langchain-ai/docs/issues/3730

Type of issue

issue / bug

Language

Python

Description

Summary

When you pass an OpenAI model string such as openai:gpt-5.4 to create_deep_agent, Deep Agents resolves it via resolve_model → init_chat_model with harness profile kwargs from libs/deepagents/deepagents/profiles/_openai.py, which registers use_responses_api=True by default.

That behavior is documented in Python docstrings (resolve_model, create_deep_agent “OpenAI Models and Data Retention” note) but not on the public Deep Agents Models page (src/oss/deepagents/models.mdx). Readers who configure models only from the docs may assume Chat Completions–style defaults or miss data-retention implications.

Library references

  • libs/deepagents/deepagents/profiles/_openai.py — _HarnessProfile(init_kwargs={"use_responses_api": True}) for provider openai.
  • libs/deepagents/deepagents/_models.py — resolve_model merges profile init_kwargs into init_chat_model; docstring states OpenAI string models default to the Responses API.
  • libs/deepagents/deepagents/graph.py — create_deep_agent docstring note on Responses vs chat completions (use_responses_api=False) and optional store / include for retention.

Expected documentation

On src/oss/deepagents/models.mdx (and optionally a one-line cross-link elsewhere):

  • State explicitly that openai: string specs default to **use_responses_api=True` when used with Deep Agents.
  • Explain how to opt into Chat Completions instead: init_chat_model("openai:...", use_responses_api=False) and pass the initialized model instance to create_deep_agent (so harness defaults for that flag are not applied the same way as for a bare string).
  • Summarize or link the data retention guidance already in the create_deep_agent docstring for Responses API users.

Why it matters

Users should not have to read source or API docstrings to discover a default that changes which OpenAI API surface is used and affects retention-related knobs.

Originally created by @AdemBoukhris457 on GitHub (Apr 25, 2026). Original GitHub issue: https://github.com/langchain-ai/docs/issues/3730 ### Type of issue issue / bug ### Language Python ### Description ## Summary When you pass an OpenAI model **string** such as `openai:gpt-5.4` to `create_deep_agent`, Deep Agents resolves it via `resolve_model` → `init_chat_model` with harness profile kwargs from `libs/deepagents/deepagents/profiles/_openai.py`, which registers `use_responses_api=True` by default. That behavior is documented in Python docstrings (`resolve_model`, `create_deep_agent` “OpenAI Models and Data Retention” note) but **not** on the public Deep Agents **Models** page (`src/oss/deepagents/models.mdx`). Readers who configure models only from the docs may assume Chat Completions–style defaults or miss data-retention implications. ## Library references - `libs/deepagents/deepagents/profiles/_openai.py` — `_HarnessProfile(init_kwargs={"use_responses_api": True})` for provider `openai`. - `libs/deepagents/deepagents/_models.py` — `resolve_model` merges profile `init_kwargs` into `init_chat_model`; docstring states OpenAI string models default to the Responses API. - `libs/deepagents/deepagents/graph.py` — `create_deep_agent` docstring note on Responses vs chat completions (`use_responses_api=False`) and optional `store` / `include` for retention. ## Expected documentation On `src/oss/deepagents/models.mdx` (and optionally a one-line cross-link elsewhere): - State explicitly that **`openai:` string specs` default to **`use_responses_api=True`** when used with Deep Agents. - Explain how to opt into **Chat Completions** instead: `init_chat_model("openai:...", use_responses_api=False)` and pass the **initialized model instance** to `create_deep_agent` (so harness defaults for that flag are not applied the same way as for a bare string). - Summarize or link the **data retention** guidance already in the `create_deep_agent` docstring for Responses API users. ## Why it matters Users should not have to read source or API docstrings to discover a default that changes which OpenAI API surface is used and affects retention-related knobs.
yindo added the externaldeepagents labels 2026-06-05 17:26:37 -04:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
bug
deepagents
docs-infra
external
external
help wanted
integration
internal
internal
javascript
js/ts
langchain
langgraph
langsmith
needs-title
pull-request
Mirrored from GitHub Pull Request
 python
reference
No Label deepagents external
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/docs#2753
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?