Danielxu0208 / rsi_eval
git clone https://github.com/Danielxu0208/rsi_eval.git
Python SDK and CLI for evaluating recursive self-improvement methods on agent benchmarks through a single evaluate() interface.
Why you might care
Lets you run your own RSI loop against existing benchmarks like Terminal-Bench, SWE-bench, and GPQA without rebuilding task execution and grading. Returns per-task scores, status, token usage, and verifier output in a common format, and snapshots candidates for resumable experiments. The core SDK has no third-party runtime dependencies.
+31
stars, 24h · #1,716
+106
stars, 7 days · #2,792
+282
stars, 30 days · #3,595
282
★ total · +60% this week
13
forks · 0 open issues
Star history (as recorded here)
LanguagePython
LicenseApache-2.0
Created2026-10-12 (20d ago)
Last push2d ago
Watchers2
OwnerUser
GitHub topics
Links
README
Loading README…
Blurb generated by claude-haiku-5-5 1d ago. Star deltas marked ~ are extrapolated from partial history.