Repository navigation
Route gh-aw#74188's workflow with economy instead of balanced - #74586
Merged
Merged
Conversation
Balanced picks gpt-5.6-sol:max for the router's implement/hard labels on every run (405-616 AIC per run vs ~55 on the old gpt-5.6-terra pin). Economy picks gpt-5.6-luna:max for the same labels. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
SivaKesava1
changed the title
Contributor
There was a problem hiding this comment.
🔵 Needs a closer look
Approval should wait for human validation because strict compilation and external security scanning were unavailable, while the selected model, AIC cost, and generated parallelization quality remain unverified.
No open issues 🎉
What changed in this PR
This pull request changes the Daily Go Test Parallelizer’s model routing from balanced to economy to reduce AI-credit usage, and regenerates its workflow lock with AWF v0.29.6.
Changes:
- Switches model routing from
balancedtoeconomy. - Regenerates the lock with updated AWF/MCPG versions and matching routing configuration.
- Preserves existing permissions, strict validation, network restrictions, draft-PR behavior, and
**/*_test.gooutput limits.
| File | Description |
|---|---|
.github/workflows/daily-go-test-parallelizer.md |
Selects economy routing while preserving workflow safety and output constraints. |
.github/workflows/daily-go-test-parallelizer.lock.yml |
Regenerated workflow lock reflecting economy routing and updated dependencies. |
Review effort: Lite (auto)
Contributor
|
🎉 This pull request is included in a new release. Release: |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Switches Daily Go Test Parallelizer's model routing from
balancedtoeconomy.Why: since it was routed in #74200 (2 Nov), the classifier labels this workflow's fixed prompt
implement/subsystem/hardorimplement/multi_file/hardon every run, andbalancedpicksgpt-5.6-sol:maxfor those cells. Its runs since then cost 405–616 AIC each (agent plus detection), against ~55 AIC per run on the oldgpt-5.6-terrapin (Ryan's rollout tracker: ~4,600 AIC/week). At 84 runs a week that is roughly 40,000 AIC/week, and every completed run ended innoop.Change:
mode: balanced→mode: economyindaily-go-test-parallelizer.md, plus the regenerated lock. In AWF v0.29.6's tables,economypicksgpt-5.6-luna:maxfor the sameimplement/*/hardlabels (the pick #74188 expected; modelled at ~1,300 AIC/week). The workflow still opens draft PRs restricted to**/*_test.go, behind tests and human review.The lock also moves from AWF 0.29.5 to 0.29.6, because
main's locks haven't been recompiled since #74389 bumped the default; recompiling only this workflow produces that.Follow-up: dispatch on this branch to confirm the pick and cost, then watch its scheduled runs.