docs(benchmarks): TB 2.1 MiniMax-M3 v0.33.0 re-run (46/89 · 51.7%)#113
Merged
Conversation
Second TB 2.1 MiniMax-M3 run, this time on a binary that includes PR #97's stream-drop auto-retry. Headline: ConnectionDropped count fell from 16 to 2 (-87.5%) and score moved +4.5 pp (47.2% → 51.7%). Resolved% slipped -4.0 pp because the recovered trials were mostly non-trivial — they enlarged the denominator faster than the numerator.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #113 +/- ##
=======================================
Coverage 74.17% 74.17%
=======================================
Files 65 65
Lines 16798 16798
=======================================
Hits 12460 12460
Misses 4338 4338 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Second TB 2.1 MiniMax-M3 baseline, this time on the binary that includes PR #97's stream-drop auto-retry. Compared against PR #98's 06-02 baseline:
ConnectionDropped)ConnectionDropped+ 1VerifierTimeoutError)TimedOut)TimedOut)PR #97 cut
ConnectionDropped16 → 2 (-87.5%) — the goal of that change. Resolved% dipped because the recovered trials weren't all easy ones (denominator grew faster than numerator).Token cost is the bill for auto-retry: replays send context again on each attempt. Cache-hit rose to 93.8%, so most of the replayed tokens were cache reads.
Files
benchmarks/RESULTS.md— new row at the top, drop "After PR feat(agent): auto-retry on stream drop with Reconnecting event #97 lands…" from Pendingbenchmarks/terminal-bench/history/tb21-minimax-m3-20260603.csv— per-trial breakdownThe Vercel HTML report is already live at https://ignis-bench-reports.vercel.app/reports/tb21-minimax-m3-20260603 (separate repo).
Test plan
aggregate_results.py→ 46p/40f/3e, sub-typesTimedOut:12 / ConnectionDropped:2 / VerifierTimeoutError:1