Skip to content

chore(db): clean up AgentX results - #723

Merged
cquil11 merged 2 commits into
masterfrom
agent/purge-unsupported-agentx-models
Aug 11, 2026
Merged

chore(db): clean up AgentX results#723
cquil11 merged 2 commits into
masterfrom
agent/purge-unsupported-agentx-models

Conversation

@cquil11

@cquil11 cquil11 commented Aug 11, 2026

Copy link
Copy Markdown
Contributor

Summary

Purge these exact production attempts:

  • Kimi K2.5 FP4 B300 vLLM MTP: run 29506569772, attempt 2 — AgentX is no longer supported for this model.
  • MiniMax M2.7 FP4 B200 SGLang MTP: run 29765418393, attempt 5 — AgentX is no longer supported for this model.
  • DeepSeek V4 FP4 B300 vLLM MTP: run 29778042138, attempt 1 — outdated AgentX harness; corrected by v1+ runs 31192604550 and 31415828111.
  • DeepSeek V4 FP4 B200 vLLM MTP: run 29778042858, attempt 2 — outdated AgentX harness; corrected by v1+ run 31192602558.
  • Qwen3.5 FP4 GB300 Dynamo-SGLang MTP: run 30391410523, attempt 1 — outdated AgentX harness; corrected by v1+ run 31042542308.
  • Kimi K3 FP4 B300 vLLM MTP: run 30425131777, attempt 3 — outdated AgentX harness.

Only the listed old attempts are purged. The corrected v1+ runs are explicitly preserved. The overrides prevent future re-ingestion and are applied automatically to production after merge.

Upstream cleanup

The four newly added outdated-harness purges do not remove their upstream configurations.

Validation

Checks were not rerun for the follow-up additions, per request. The original change passed the purge override tests, lint, formatting, and typecheck.


Note

Medium Risk
Deletes specific production benchmark attempts via the override pipeline. Risk is limited to wrong run/attempt IDs being purged, with CI verification after merge.

Overview
Adds 6 purge overrides to PURGED_RUN_ATTEMPTS so specific AgentX benchmark attempts are deleted from production and blocked from re-ingest.

Two attempts are purged because AgentX is no longer supported for those models (Kimi K2.5, MiniMax M2.7). Four more are purged for an outdated AgentX harness, with several noting corrected replacement runs.

Reviewed by Cursor Bugbot for commit d823952. Bugbot is set up for automated code reviews on this repo. Configure here.

Purge Kimi K2.5 run 29506569772 attempt 2 and MiniMax M2.7 run 29765418393 attempt 5 because AgentX is no longer supported for those models.
@vercel

vercel Bot commented Aug 11, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
inferencemax-app Ready Ready Preview Aug 11, 2026 8:39pm

Request Review

@cquil11 cquil11 changed the title chore(db): clean up unsupported AgentX results chore(db): clean up AgentX results Aug 11, 2026
@cquil11
cquil11 merged commit da43592 into master Aug 11, 2026
24 checks passed
@cquil11
cquil11 deleted the agent/purge-unsupported-agentx-models branch August 11, 2026 20:40
chunfangamd added a commit to chunfangamd/InferenceX-app that referenced this pull request Aug 12, 2026
Pulls in SemiAnalysisAI#722 and SemiAnalysisAI#723, two ETL run-override additions that exclude
unsupported AgentX results. No conflicts: they touch only
packages/db/src/etl/run-overrides.ts and share no file with the overview work
on this branch.

Co-authored-by: Cursor <cursoragent@cursor.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant