LetTokenBurn$ npm i -g lettokenburn-cli$ ltb bench run swe-rebench/swe-rebench-leaderboard@2 --benchmark swe-rebench -a claude-code -m anthropic/claude-opus-4-8$ ltb bench push
Harbor-native: you can also call Harbor directly.
$ harbor run -d swe-rebench/swe-rebench-leaderboard@2 -a claude-code -m anthropic/claude-opus-4-8
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
No sourced scores recorded yet.
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
SWE-Rebench (SWE Re-benchmark) — 重新设计的SWE任务. Scoring: Resolved%. Difficulty: 极高. Imported from benchmark_research.md (2026-06).