Refresh harness leaderboard snapshots#40
Open
evanatpizzarobot wants to merge 1 commit into
Open
Conversation
Updated lastUpdated to 2026-06-15. Benchmark scores unchanged: official upstream sites (swebench.com, tbench.ai, aider.chat) returned HTTP 403 during this run so per-entry scores could not be re-verified. The Aider Polyglot YAML on GitHub is current only through October 2025 and does not yet include our tracked model names (Claude Opus 4.7, GPT-5.5, DeepSeek V4 Pro). Added a one-line TODO in note for Kilo Code, the official successor to roo-code which shut down 2026-05-15. https://claude.ai/code/session_01SVZWBi9UTPJ1VrrBUnKda9
Deploying tensorfeed with
|
| Latest commit: |
5a67f9a
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://c3cd488e.tensorfeed.pages.dev |
| Branch Preview URL: | https://harness-refresh-2026-06-15.tensorfeed.pages.dev |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
lastUpdatedbumped to2026-06-15in bothdata/harnesses.jsonandworker/src/harnesses.ts(in sync, byte-for-byte on the data fields).notefield: Roo Code shut down on 2026-05-15 and Kilo Code (kilocode.ai) is the official successor. A human needs to decide whether to add Kilo Code and updatesrc/lib/harness-directory.ts.Access Issues Encountered
The three canonical upstream sources blocked automated fetching:
swebench.com(SWE-bench Verified): HTTP 403tbench.ai(Terminal-Bench): HTTP 403aider.chat/docs/leaderboards/(Aider Polyglot): HTTP 403More than 15 third-party aggregator and mirror sites also returned 403. The one source that was accessible was the Aider Polyglot YAML on GitHub (
Aider-AI/aiderraw file), but its most recent entries date to October 2025 and do not include the model names tracked in our data (Claude Opus 4.7, GPT-5.5, DeepSeek V4 Pro). Scores for those entries cannot be updated until the Aider team publishes results for those models.SWE-Lancer: The original benchmark repository (
openai/SWELancer-Benchmark) has been archived and merged intoopenai/preparedness. No new public leaderboard with per-agent scores was found.Scores Changed
None. Per-entry scores are unchanged from the 2026-04-30 snapshot.
TODOs Added to Note
roo-code shut down 2026-05-15; Kilo Code (kilocode.ai) is the official successor and should be evaluated for addition to harness-directory.ts after editorial review.Test Plan
npx tsc --noEmitpasses at repo root (afternpm install)cd worker && npx tsc --noEmitpasses (afternpm install)https://claude.ai/code/session_01SVZWBi9UTPJ1VrrBUnKda9
Generated by Claude Code