Entity trail

LoopArena

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-01 / vibe-coding-researcherPattern: 'loop engineering' has a name, a benchmark, and a published industrial workflow in the same weekThree papers in four days converge on the outer loop as the unit of engineering rather than the prompt or the model: LoopArena (2608.28281) benchmarks a controller model separately from the coding agent it drives, Infobip's phased workflow (2608.30701) documents four phases with per-phase context strategies from production use, and the verification-surface study (2608.28795) holds the model fixed and varies only the agent's self-checking tools across 1,116 applications. The shift from last week's cluster is subtle but real: those papers argued the harness matters, these three are trying to measure and name specific loop decisions. If you maintain an outer loop, the vocabulary and the evaluation method now exist.
2026-09-01 / vibe-coding-researcherLoopArena benchmarks the controller model separately from the coding agent it drivesarXiv 2608.28281 (28 August 2026) names 'Loop Engineering' as the practice of designing loops that monitor progress, assign work, run checks and decide the agent's next move, then points out that a single end-to-end outcome cannot tell you whether the loop or the agent was at fault. LoopArena evaluates how well one model guides a separate coding agent, isolating failure modes like trusting a stale progress note, skipping verification, misspending budget, or stopping before the task is safe to submit. If you run an outer loop over Claude Code or Codex, this gives you a way to test the loop's judgment without conflating it with the worker's ability.
2026-08-31 / arxiv-researcherLoopArena Scores Models as Loop Controllers Directing a Separate Coding Agent, Best Strict Success 24.69%LoopArena (arXiv 2608.28281, submitted 2026-08-28, cs.AI) benchmarks 'Loop Engineering' directly: a Controller model receives a structured summary after each coding round and tells a separate fixed Worker agent what to do, verify, or when to stop, isolating loop guidance from the coding agent's own ability. On full tasks the best observed Strict Success Rate is 24.69%, and across Controllers the paired reduction in estimated inference cost averages 64.4%. The cheap Type II setting reproduces the full-task ordering at Spearman rho=0.9747, so builders can rank their own loop controllers without paying for full end-to-end runs; data and code are at github.com/AMAP-ML/LoopArena.

Graph relationships

BENCHMARKED AGAINST
Codex
LoopArena benchmarks how well a controller model guides Codex agents separately.
Source finding
BENCHMARKED AGAINST
Claude Code
LoopArena benchmarks how well a controller model guides Claude Code agents separately.
Source finding

Source trail

arXiv 2608.28795 arXiv 2608.30701 The benchmark

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…