Fetching from the wire…
01
02
03
04
05
06
07
08
Research2026-07-29 · source-backed
arXiv 2607.25886 isolates data-centric research capability by fixing the entire post-training stack so only the agent's data strategy varies. Four frontier agents across six benchmarks. Among searches that continued past the best observed score, 78.26% ended on a lower-scoring final attempt and the rest merely recovered the peak. The practical read: your self-improvement loop needs explicit checkpoint preservation, not more iterations. Code at github.com/evolvent-ai/RSIBench-Data.
Each link below shares sources, entities, or timing with this story.
Shared entity: Code / Same source domain / Shared topic / Earlier coverage / Tension
Both cover Code; reported by the same outlet (arxiv.org); overlapping topics (agent, benchmark, code).
Shared entity: Data / Same source domain / Shared topic / Earlier coverage / Downstream implication
Both cover Data; reported by the same outlet (arxiv.org); overlapping topics (capability, data).
Shared entity: Code / Same source domain / Shared topic / Earlier coverage
Both cover Code; reported by the same outlet (arxiv.org); overlapping topics (agent, code, their).
Both cover Code; reported by the same outlet (arxiv.org); overlapping topics (agent, benchmark, code).
Both cover Code; reported by the same outlet (arxiv.org); overlapping topics (agent, capability, code).
Shared entity: Among / Shared topic / Earlier coverage / Tension
Both cover Among; overlapping topics (agent, among, code); earlier Among coverage from 2026-03-16.
Shared entity: Code / Same source domain / Shared topic / Earlier coverage
Both cover Code; reported by the same outlet (arxiv.org); overlapping topics (agent, code).
Both cover Code; reported by the same outlet (arxiv.org); overlapping topics (benchmark, code).