{
  "id": 16100,
  "url": "https://arxiv.org/abs/2608.02391v1",
  "title": "Cooperative Coevolution for Resource-Constrained Agentic LLM Post-Training",
  "summary": "Tool-using large language model (LLM) agents produce long, multi-turn trajectories, making gradient-based post-training memory-intensive. Evolution strategies (ES) enable memory-efficient full-parameter post-training without backpropagation and can eventually match the performance of gradient-based reinforcement learning (RL). However, resource-constrained settings typically offer only a few GPUs, so the high GPU-hour requirements of ES translate into prohibitively long training times. To addres",
  "authors": "Zhiyuan Wang, Shengcai Liu, Jiahao Wu, Ning Lu, Hui Ouyang, Shaofeng Zhang, Haoze Lv, Ke Tang",
  "category": "research",
  "topics": "agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-03T15:34:45.000Z",
  "fetched_at": "2026-08-04T05:10:21.797Z",
  "source_slug": "x-arxiv-cs-ai",
  "source_name": "arXiv cs.AI",
  "source_homepage": "https://arxiv.org/list/cs.AI/recent",
  "ethics_ai_record_url": "https://ethics.ai/record/16100",
  "original_url": "https://arxiv.org/abs/2608.02391v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}