Repository navigation
MAINT: Support OpenAI 2 and 3 - #3214
Merged
Merged
Conversation
dependabot
Bot
added
dependencies
Contributor
dependabot
Bot
force-pushed
the
dependabot/uv/openai-3.29.0
branch
from
October 15, 2026 02:55
256bc34 to
e57cd44
Compare
Contributor
Bumps [openai](https://github.com/openai/openai-python) from 2.54.0 to 3.29.0. - [Release notes](https://github.com/openai/openai-python/releases) - [Changelog](https://github.com/openai/openai-python/blob/main/CHANGELOG.md) - [Commits](openai/openai-python@v2.54.0...v3.29.0) --- updated-dependencies: - dependency-name: openai dependency-version: 3.29.0 dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com>
dependabot
Bot
force-pushed
the
dependabot/uv/openai-3.29.0
branch
from
October 15, 2026 03:57
e57cd44 to
9f9f21d
Compare
Bound OpenAI to >=2.20.0,<4 and temporarily pin the SDK-compatible LiteLLM development release. Cover native and legacy transports, errors, authentication, binary responses, embeddings, and realtime on both SDK lines, with a CI matrix for base and LiteLLM installations. Document the stable LiteLLM release gate and retain compatible video endpoints with narrow deprecation exceptions. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 963182a0-189b-4a74-bbbd-f7229781f6e2
Roman Lutz (romanlutz)
changed the title
Contributor
Contributor
Contributor
Contributor
Contributor
Contributor
Restore the stable LiteLLM and OpenAI 2 development lock. Test SDK 3 in base installations without LiteLLM, restrict stable LiteLLM compatibility jobs to SDK 2, and remove the unnecessary prerelease release gate. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 963182a0-189b-4a74-bbbd-f7229781f6e2
Roman Lutz (romanlutz)
approved these changes
Oct 15, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Support OpenAI SDK 2 and 3 without requiring a LiteLLM prerelease.
openai>=2.20.0,<4. SDK 2 remains supported, while default package installations can select SDK 3. The next untested major version is excluded.openai<3requirement selects the supported SDK 2 line when that extra is installed. No third-party locked package differs from main.Tests and Documentation
Current head
21b6f405: all 36 required checks passed. build_and_test passed all 24 OS/Python/extras test jobs, all eight SDK compatibility jobs, and all three pre-commit jobs. Coverage, Docker, frontend, documentation, and CodeQL workflows also passed. The previously failing macOS 3.11/all-extras job passed on this head without skipping tests or weakening analytics deadlines.Local environment: Windows x64, Python 3.14.4. Provider responses and WebSocket frames are fabricated; no provider credentials or paid API calls were used.
uv run --frozen --no-sync python -m pytest -q -n 4 --dist=loadfile --timeout=120 tests\unit\prompt_target tests\unit\embedding tests\unit\auth tests\unit\backend\test_analytics_routes.py tests\unit\memory\test_attack_analytics_eval_identity.py --tb=short: 2,256 passed, 1 skipped with stable LiteLLM 1.104.0 and OpenAI 2.54.0, including both analytics files from the old macOS failure.uv run --frozen --no-sync --with openai==2.20.0 python -m pytest -q --timeout=120 tests\unit\prompt_target\target\test_openai_sdk_compatibility.py tests\unit\prompt_target\target\test_openai_url_warnings.py tests\unit\prompt_target\target\test_litellm_chat_target.py --tb=short: 177 passed with stable LiteLLM.uv sync --frozenverified that LiteLLM was absent: compatibility, URL, error-handling, target-auth, and embedding tests each produced 159 passed, 8 optional LiteLLM cases skipped on SDK 3.0.0 and 3.29.0. The full commands are encoded in the compatibility workflow and were also exercised by the passing current-head matrix.uv run --frozen --no-sync ty check pyrit tests\unit\prompt_target\target\test_openai_sdk_compatibility.py tests\unit\prompt_target\target\test_openai_url_warnings.py --output-format concise: passed with stable dependencies and all extras.uv run --frozen --no-sync python -m build_scripts.validate_docs: passed.uv lock --check --offline: passed.Documentation: updated
doc/code/targets/0_prompt_targets.mdand OpenAI target constructor documentation. No notebook code changed, so JupyText execution was not needed. Live-provider integration tests were not run.