docs(release_notes): add v1.94.0rc1 release notes - #609
Merged
Conversation
236 content PRs across the range, the largest release since the 1.93 line.
Headline items: the Router(plugins=[...]) pipeline and Auto-Router v2, MCP
client-held credentials (dcr_bridge oauth_delegate SSO, gateway-bound envelopes,
issuer-anchored discovery, ID-JAG egress), the shared DataTable dashboard
migration, Python 3.14 support, and per-model prompt cache minimums.
Predecessor resolution is the part worth reviewing. The obvious pick,
v1.93.0-rc.2, FAILS the ancestry check -- rc.2 tags are cut from release
branches, not main, so it is not an ancestor of v1.94.0-rc.1. The correct
predecessor is v1.93.0-rc.1, the most recent RC that is genuinely an ancestor:
v1.93.0-rc.2 2026-07-15 not-ancestor (cut from the 1.93.x branch)
v1.93.0-rc.1 2026-07-11 ANCESTOR -> 353 non-merge commits
PRs were collected from merge commits rather than squashed-commit refs: this
repo lands PRs as "Merge pull request #N", so 173 of the 353 commits carry no
(#N) of their own. Union of 60 merge-landed and 181 squash-landed PRs, minus 5
litellm_internal_staging vehicle PRs (#33308, #33425, #33491, #33640, #33868),
gives the 236 content PRs.
167 are bulleted; the remaining 69 CI / chore / test / build / version-bump PRs
are counted in the ownership roll-up but not bulleted, matching v1.93.0rc1.
167 + 69 = 236 reconciles exactly, and no cited PR falls outside the source set.
Verified rather than assumed:
- Docker tag probed on GHCR: 1.94.0-rc.1 -> 200; 1.94.0rc1 and every main-v*
form -> 404, so the Deploy block uses the hyphenated form.
- pip: litellm==1.94.0rc1 resolves on PyPI.
- New contributors: all 16 authors checked against their first-ever merged PR.
Exactly one is new (@Napuh, #32667); the GH release body was not trusted.
- Compiles with @mdx-js/mdx, and no bare `<` outside code spans -- the failure
mode that broke the v1.93.0 build on an unescaped `<3.15`.
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
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.
236 content PRs — the largest release since the 1.93 line.
Headline items: the
Router(plugins=[...])pipeline and Auto-Router v2 · MCP client-held credentials (dcr_bridgeoauth_delegateSSO, gateway-bound envelopes, issuer-anchored discovery, ID-JAG egress) · shared DataTable dashboard migration · Python 3.14 support · per-model prompt cache minimums.Predecessor resolution — the part worth reviewing
The obvious pick, v1.93.0-rc.2, fails the ancestry check. rc.2 tags are cut from release branches, not
main, so it is not an ancestor of v1.94.0-rc.1:Comparing against rc.2 would have produced a diff mixing two branch lines.
PR collection
This repo lands PRs as
Merge pull request #N, so 173 of the 353 commits carry no(#N)of their own — collecting from squashed-commit refs alone would have missed most of the release. Final set is the union of 60 merge-landed + 181 squash-landed PRs, minus 5litellm_internal_stagingvehicle PRs (#33308, #33425, #33491, #33640, #33868) = 236.167 bulleted; 69 CI/chore/test/build/version-bump PRs counted in the ownership roll-up but not bulleted, matching how v1.93.0rc1 handles its
Other: 56bucket. 167 + 69 = 236 reconciles exactly, and no cited PR falls outside the source set.Verified, not assumed
1.94.0-rc.1→ 200;1.94.0rc1and everymain-v*form → 404. Deploy block uses the hyphenated form.litellm==1.94.0rc1resolves on PyPI.@mdx-js/mdx, with no bare<outside code spans — the exact failure that broke the v1.93.0 build on an unescaped<3.15(fixed in docs(release_notes): promote v1.93.0 to stable #603).Judgement calls
A handful of PRs sit on section boundaries — e.g. #31592 (OTLP billable-request metering) is filed under Logging but is arguably Spend Tracking. Happy to move any of them.