{"arms":"fresh_prior_exact_id_only_no_overlap_no_feedback","availability":"Pool histórico de referencia; no afirma disponibilidad o elegibilidad del pool productivo actual","baselines":["claude-3-haiku-20240307","gemini-2.0-flash-001","gpt-4o-mini"],"cache_sha256":{"claude-3-haiku-20240307":"9f5bed666a524d74e629091b57ef2fdf644181dcc1039ef44b44460f254d150a","gemini-2.0-flash-001":"5c2364f65a4c309bc531338672445894957a6465be71e31a21a6727028404c20","gpt-4o-mini":"84c83a87977dc1bab556aee6b288ad8a8f17bee4ede80bba29f235c8d29a2e4e"},"configuration":{"frozen_hour":10,"objective":"balanced","plan_id":"pro-code"},"corpus":{"name":"RouterArena · sub_10 integrado","revision":"cff9659dd6a3a07815f4842fcdadea5a597a1df1","selection":"Todos los 809 IDs de Claude, iguales a Gemini; GPT mini se filtra a esos IDs; orden de Claude; igualdad exacta de preguntas","source_url":"https://github.com/RouteWorks/RouterArena/tree/cff9659dd6a3a07815f4842fcdadea5a597a1df1","total":809},"cost":"Suma de evaluation_result.inference_cost cacheado en USD; cero gasto nuevo; TT sólo es peso interno, no precio USD","created_at":"2026-09-25T11:53:36.444460+00:00","expected_weights_tt":{"claude-3-haiku-20240307":1,"gemini-2.0-flash-001":1,"gpt-4o-mini":1},"experiment":"titan-u-routerarena-sub10-historical-v1","inferences":0,"inputs_sha256":"88444b587767b8806dfdc249ac8bb8b695a95d24774209b4d37556379f17efcb","learning":"frozen_no_updates","limitations":["Subconjunto integrado sub_10 de 809 casos; no es score oficial completo ni ranking de RouterArena","Los tres arms son frescos; sin entrenamiento o ajuste con el conjunto","Costes y calidad corresponden a respuestas históricas cacheadas, no a inferencias de TITAN","Latencia mide sólo el selector local"],"missing_data":"Cualquier ausencia, duplicado, pregunta diferente, coste inválido o score inválido hace fallar la ejecución completa","model_ids":["claude-3-haiku-20240307","gemini-2.0-flash-001","gpt-4o-mini"],"oracle":null,"phases":["prepare sin etiquetas","predict autorizado tras prerregistro","evaluate autorizado tras sello completo de predicciones"],"pool_sha256":"58487ec54880fb2a2b5391f348130c20756cc37262113d2bee830bbf28d53e3f","publication":"Sólo protocolo, IDs/hashes, decisiones y agregados; ninguna pregunta/respuesta/caché redistribuida","quality":"Media aritmética de evaluation_result.score cacheado en [0,1], sin transformar; correct=null; no se presupone clasificación binaria","router_input":"Sólo question: sin IDs, etiquetas, respuestas, costes por caso, eval_name ni resultados","runtime":{"machine":"arm64","python":"3.9.6","system":"Darwin"},"sampling":"none","schema_version":1,"scope":"internal_offline_historical_reference_not_official_arena_score","snapshot_sha256":"dd7ce8026b1a8b31a9edd339a7c98f046e7b85779b6103956fbb68b3297aa3a8","source_sha256":{"scripts/benchmarks/titan_cached/__init__.py":"6d96d481096aa7b68840f5014cecf1e1d6521d5d1c6bf596ef55959c7a582d9d","scripts/benchmarks/titan_cached/frozen.py":"d06c1247d3ec28f2d8a3463b75ecd0352bea080eb32961293e1015f2baa10f75","scripts/benchmarks/titan_routerarena/__init__.py":"8fb635e6afa8ff2bfa17ab03105d1762f0b69a66d4e6fd3f6f0a5c033fe5965d","scripts/benchmarks/titan_routerarena/run.py":"d8547e602a5931483c3ab63693febe932b51c4ecb178eb9263072a427b93e755","scripts/benchmarks/titan_routing_audit/__init__.py":"6eb8ea5976a48e11b73afa0cc1e8727bf97497255262e38ef14906f8228cb27c","scripts/benchmarks/titan_routing_audit/run.py":"8679a7787aab044b1e5f61a4ecb04ff2f72344d6ddebdfadfa174864f4d4845a","server/native_approvals.py":"f35d727d074419d0ef5c21802b9ae82ca746c71364be31f5930813042e4022e3","server/theus_oauth_server.py":"fb318a5f8c4bacfec40e15b0057c49fc01e8e9245924ee6ffc1d19f7ca8e0a13","server/titan_contract.py":"6499a5a8bd102eec530ff5c93fbf5e47419f81bd150a58b4c02867c7b30c9004","server/titan_evidence.py":"997d5deb99cf43ea8d2ef29e7477fe50f8c61cee8c5196aacc9f82457877b742","server/titan_evidence_http.py":"814b3b53b8357c8cc0375d0d83d7c2e07374b9ab83b259f285b37df8f63e4fad","server/titan_gateway_adapter.py":"b13dc2e80f29bb8f15d3e8999df2714926adad0b9fd8b837a83804e34fa97c74"},"static_prices_sha256":"99cc29b3e37eef0906ce1c5ff682149ba9d306692fe4a4693d49784fcbd778e0","training":"disabled"}
