Conversation
Google released gemini-3.7-flash. Measured against the shipped serving pin gemini-3.6-flash on the identical apparatus (22 live-quality fixtures, fixed judge deepseek-chat thinking-off, both candidates in one session): 19 pass / 2 warn / 1 error vs 18 / 4 / 0, half the ai_not_improved warns, ~2x faster (4.2s vs 8.9s/rewrite), identical published pricing. Allowlist only — every pinned default including the serving pin is untouched: 3.7 shows a reproducible MPS-50 meaning-loss tail on en-social-01 (2 of 4 samples) below the incumbent's 75 floor, and worst-case MPS is the deciding column. Evidence: docs/operations/serving-engine-gemini-3.7-flash-20260813.md. v6.4 hold re-frozen for the mutated contract test; changelog historical pins shifted +1 for the new Unreleased bullet. 1649 tests, lint, preflight-hold validator, retired-concepts scanner, release:check all pass.
Merging per repo policy (agent-owned review+merge into dev). Independent review executed as a mechanical checklist after both subagent review lanes failed on exhausted model backends: (1) no preset default or PATINA_PRO_*/FREE_* runtime pin moved — all diff mentions are docs/comments; (2) playground contract copy byte-identical; (3) v6.4 re-frozen hash matches sha256 of the mutated test file in both validator and checked-in state; (4) retired-concepts diff changes line numbers only, all 13 hashes intact; (5) evidence doc discloses the deepseek-judge caveat and the en-social-01 MPS-50 tail, serving pin unchanged. Full gate: 1649 tests / lint / hold validator / scanner / release:check all pass.
feat: add auditable conversion and measured rewrite controls
chore(release): prepare 7.1.0
Contributor
|
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.
Patina 7.1.0
Production release via merge from
devtomain.Included
offby default)Verification
npm run release:checknpm run lintnpm test: 1683 pass, 2 skippedMerge commit required; do not squash.