Yashp2003's picture
Update logbook: Reproduction: SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?
bd3727b verified
|
Raw
History Blame
2.46 kB

Claim 1: SWE-Bench Pro comprises 1,865 problems sourced from 41 actively maintained software repositories (abstract only).


Claim 1 — Scale of the benchmark

Claim (verbatim): "SWE-Bench Pro comprises 1,865 problems sourced from 41 actively maintained software repositories."

Verdict: Partially verified — the public subset (731 problems / 11 repos) is directly countable; the full 1,865 / 41 figure is stated by the paper and corroborated by the official release partner Scale AI.

What we verified

We loaded the official public benchmark release ScaleAI/SWE-bench_Pro (the canonical HF copy released by Scale AI, the paper's benchmark partner) and counted its contents:

  • 731 problems across 11 repositories in the publicly accessible split.
  • Per-repository instance counts (all 11 public repos):
Repository Instances
ansible/ansible 96
internetarchive/openlibrary 91
flipt-io/flipt 85
qutebrowser/qutebrowser 79
gravitational/teleport 76
future-architect/vuls 62
protonmail/webclients 65
navidrome/navidrome 57
element-hq/element-web 56
NodeBB/NodeBB 44
tutao/tutanota 20

Relationship to the headline 1,865 / 41

The paper states the full benchmark is 1,865 problems from 41 repos, partitioned into a public set of 11 repos, a held-out set of 12 repos, and a commercial set of 18 proprietary repos (11 + 12 + 18 = 41). The 11 public repos we counted exactly match the "public set of 11" in Claim 2, and 731 is the released count of that public set. The held-out (12) and commercial (18) splits are explicitly not publicly accessible by design, so their 1,134 instances (1,865 − 731) cannot be independently counted — this is consistent with the paper's contamination-resistance design (Claim 5) rather than a discrepancy.

Evidence