{"slug": "cost-effective-agent-harnesses-for-abstract-reasoning-and-generalization-on-arc", "title": "Cost-Effective Agent Harnesses for Abstract Reasoning and Generalization on ARC-AGI-1", "summary": "Researchers developed cost-effective agent harnesses for the ARC-AGI-1 abstract reasoning benchmark, achieving 67.25% pass@2 at $0.62 per task using an open-weight model without fine-tuning. The Reflective Orchestrator architecture lifts a 15.50% baseline by ~52 points, demonstrating that significant improvement requires broader generation rather than better ranking.", "body_md": "arXiv:2607.06764v1 Announce Type: new\nAbstract: Recent progress on ARC-AGI-1 from disclosed architectures has come broadly from two regimes: heavy test-time compute over frontier models (evolutionary search, exhaustive sampling, extended chain-of-thought), or benchmark-specific training in which small models are fine-tuned on ARC data, often with task-specialized architectures. We study a third regime: an open-weight model in non-thinking mode (DeepSeek V3.2) under a strict budget, with no ARC-specific fine-tuning. We study what is recoverable through architecture alone, building agentic harnesses that decompose pattern-discovery and program-synthesis stages explicitly. First, we introduce an Explorer-Definer Pipeline that separates pattern discovery from executable transformation synthesis, implemented as a two-stage agent pipeline. Next, we present the Reflective Orchestrator, which augments the pipeline with autonomous exploration of new transformations when previous hypotheses fail on training pairs. On the ARC-AGI-1 public 400-task evaluation set, the pipeline reaches 57.50% pass@2 at \\$0.25 per task, and the orchestrator reaches 67.25% pass@2 at \\$0.62 per task. Together these architectures lift a 15.50% one-shot baseline by ~52 points without benchmark-specific training or heavy test-time compute. Furthermore, the orchestrator-driven lift tests a falsifiable diagnostic the pipeline produces; unbiased pass@k analysis suggests the pipeline is generation-bound, not selection-bound (selection via training-pair accuracy captures ~95% of the candidate ceiling) and predicts that significant improvement requires broader generation, not better ranking. The orchestrator implements this prediction via adaptive re-exploration and confirms it (unbiased pass@1 lift +9.81 pp, matching selection-mediated pass@2 lift). An additional pipeline ablation identifies its think tool as a significant component, with removal reducing pass@2 by 5.75 pp.", "url": "https://wpnews.pro/news/cost-effective-agent-harnesses-for-abstract-reasoning-and-generalization-on-arc", "canonical_source": "https://arxiv.org/abs/2607.06764", "published_at": "2026-07-09 04:00:00+00:00", "updated_at": "2026-07-09 04:16:58.948193+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-research"], "entities": ["DeepSeek", "ARC-AGI-1"], "alternates": {"html": "https://wpnews.pro/news/cost-effective-agent-harnesses-for-abstract-reasoning-and-generalization-on-arc", "markdown": "https://wpnews.pro/news/cost-effective-agent-harnesses-for-abstract-reasoning-and-generalization-on-arc.md", "text": "https://wpnews.pro/news/cost-effective-agent-harnesses-for-abstract-reasoning-and-generalization-on-arc.txt", "jsonld": "https://wpnews.pro/news/cost-effective-agent-harnesses-for-abstract-reasoning-and-generalization-on-arc.jsonld"}}