[KG] Claude Sonnet 4.6 — risk

{
  "entity_slug": "claude-sonnet-4-6",
  "entity_name": "Claude Sonnet 4.6",
  "entity_type": "ai_model",
  "title": "Claude Sonnet 4.6: Hitting the Human Baseline, But For How Long?",
  "narrative": "Anthropic's Claude Sonnet 4.6 sits exactly at the human OSWorld-Verified baseline of 72.1%, a notable benchmark achievement. Developed by Anthropic, it deploys Chain-of-Thought Prompting and Constitutional AI. However, the model's deployment velocity is tepid—only 3 mentions in the last 30 days. Recent news reveals a critical weakness: Anthropic's own research shows AI agents, presumably including Sonnet 4.6, failed to retrieve 261 Ebola sequences in a biology retrieval task. The model is used by King's College London, Navox Agents, and Claude Code, but faces pressure from newer adaptive thinking budgets (deprecated fixed budgets as of May 2026). The question is whether Sonnet 4.6 can maintain its baseline parity as competitors push beyond human-level performance.",
  "key_points": [
    "Scores 72.1% on OSWorld-Verified, matching the human baseline.",
    "Deploys Chain-of-Thought Prompting and Constitutional AI.",
    "Recent research reveals failure in biology retrieval (missed 261 Ebola sequences).",
    "Low mention velocity: 3 mentions in 30 days.",
    "Used by King's College London, Navox Agents, and Claude Code."
  ],
  "angle": "risk",
  "neighborhood_size": 6,
  "generated_at": "2026-06-12T15:41:13.294864+00:00"
}