{
  "id": 12772259,
  "title": "Before They Can Solve: Predicting Post-Training Coding-Agent Performance from Base Models",
  "url": "https://urgent.news/2026/10/07/before-they-can-solve-predicting-post-training-coding-agent",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-07T17:37:19.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2610.10478v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "How can we predict which base checkpoint is worth an expensive round of agentic post-training? End-to-end pass@$K$ tests whether successful behavior already appears in a base model's distribution, but it is a poor fit for agentic coding: many base checkpoints cannot reliably produce the well-formed tool invocation required to complete a task end-to-end. Single-shot or short-horizon tasks avoid…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}