Repository navigation
Repoint gemini to 2.5-flash (litellm bug); harden timeout test - #657
Merged
Merged
Conversation
… models Two fixes from the 2026-08-08 full-suite runs: 1. gemini-3.6-flash fails rdb's loop 4-of-6 times today on the litellm thought-signature converter bug (issue 654; upstream litellm 16893 / 25322 family — no fixed release as of 1.95.0/1.96.0rc1). Per the alias admission test, the gemini trio repoints to gemini-2.5-flash — the best Gemini that actually runs the loop (it predates thought signatures entirely and passed every May e2e window). Restore 3.6-flash when litellm's handling stabilizes. 2. The timeout test was outrun a second time: a model scripted the 50-file task in one bash call inside the 30s watchdog window. Third hardening makes wall-clock the constraint instead of reasoning speed: 200 files in 200 individual git commits (~40-60s of git execution alone), which no model can compress below the watchdog regardless of how fast it thinks. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Two fixes to make the 1.0 gate deterministic, from today's two full-suite runs.
Gemini: repoint to 2.5-flash pending upstream
gemini-3.6-flash failed 4 of 6 runs today on the litellm thought-signature converter bug — including 3-of-3 in the latest suite, so "intermittent" undersells it. Local bisection (1.94.2 / 1.95.0 / 1.96.0rc1 / 1.97.0.dev2) shows no fixed release, and a synthetic repro doesn't trigger it — it needs real signature flows, which is upstream's problem to chase (tracked in #654 with today's escalation). Per the admission rule — an alias points at the best model that actually runs rdb's loop — the gemini trio drops to
gemini-2.5-flash, which predates thought signatures entirely and passed every May e2e window. Restoring 3.6-flash when litellm stabilizes is a one-line repoint.Timeout test: make wall-clock the constraint
The watchdog test was outrun for the second time: the previous hardening (50 fully-specified files, 30s timeout) assumed file creation takes time — a fast model scripted all 50 in a single bash call and finished inside the window. Third design: 200 files in 200 individual git commits. Tool-execution wall time (~40-60s of git operations alone) exceeds the watchdog regardless of model speed — the constraint no longer races against reasoning latency, which only ever gets faster.
Test plan
🤖 Generated with Claude Code