{"paper":{"title":"HarnessPro: Recursive Harness Self-Improvement with Pareto Optimization for Performance-Cost Trade-offs","arxiv_id":"2610.26770","primary_cat":"cs.LG","submitted_at":"2026-10-19T17:27:38+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"Harness development plays a crucial role in building capable AI agents.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.04,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.02,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.62,"id":"s1","own":false,"role":"background","text":"Recent approaches improve harnesses through recursive self-improvement (RSI), but their performance gains often come with higher execution costs.","hedge":{"p":0.9,"label":"asserted"},"own_p":0.05,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.31,"role_probabilities":{"gap":0.38,"method":0.0,"result":0.0,"background":0.62,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":1.0,"id":"s2","own":true,"role":"contribution","text":"We introduce HarnessPro, a label-free method for evolving harnesses to improve task performance while reducing execution cost.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.87,"main_claim_p":0.26,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":1.0,"interpretation":0.0}},{"i":3,"p":0.99,"id":"s3","own":true,"role":"method","text":"In this setting, candidate harnesses are evaluated using LLM-generated tests without access to ground-truth solutions or official tests.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.6,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.99,"result":0.0,"background":0.0,"limitation":0.01,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":4,"p":1.0,"id":"s4","own":true,"role":"method","text":"Specialized agents diagnose execution trajectories, plan updates with different performance-cost priorities, and implement them as candidate harnesses.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.73,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":1.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":5,"p":0.86,"id":"s5","own":true,"role":"method","text":"Rather than returning a single harness, HarnessPro uses normalized weighted squared-distance selection to provide cost-focused (lite), balanced (medium), and performance-focused (heavy) harnesses for different deployment budgets.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.92,"main_claim_p":0.16,"quantitative":false,"quantitative_p":0.26,"role_probabilities":{"gap":0.0,"method":0.86,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.14,"interpretation":0.0}},{"i":6,"p":1.0,"id":"s6","own":true,"role":"method","text":"Selected harnesses enter the next iteration, where execution feedback guides further updates.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.62,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.03,"role_probabilities":{"gap":0.0,"method":1.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":7,"p":0.96,"id":"s7","own":true,"role":"method","text":"Our experiments cover Terminal-Bench 2, SWE-Bench Pro, and ProgramBench, using GPT-5.5 with Codex and GLM-5.2 with Claude Code.","hedge":{"p":0.97,"label":"asserted"},"own_p":0.92,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.16,"role_probabilities":{"gap":0.0,"method":0.96,"result":0.01,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.03,"interpretation":0.0}},{"i":8,"p":1.0,"id":"s8","own":true,"role":"result","text":"Across these benchmarks using GPT-5.5 with Codex, the heavy mode of HarnessPro achieves an average pass-rate gain of 2.0 percentage points and an average per-task cost reduction of 26.1% relative to RHO.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.41,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":9,"p":0.99,"id":"s9","own":true,"role":"result","text":"This advantage also holds when using GLM-5.2 with Claude Code.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.89,"main_claim_p":0.1,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.99,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.01}},{"i":10,"p":0.65,"id":"s10","own":false,"role":"contribution","text":"Code is available at https://github.com/pzs19/HarnessPro","hedge":{"p":0.51,"label":"asserted"},"own_p":0.47,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.03,"role_probabilities":{"gap":0.0,"method":0.12,"result":0.02,"background":0.04,"limitation":0.0,"future_work":0.17,"contribution":0.65,"interpretation":0.0}}]},"compiled":{"main_claim":{"p":0.86,"text":"Across these benchmarks using GPT-5.5 with Codex, the heavy mode of HarnessPro achieves an average pass-rate gain of 2.0 percentage points and an average per-task cost reduction of 26.1% relative to RHO.","location":"abstract s8","hedge_strength":{"p":1.0,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":8150,"cost_usd_reported":0.0},"compiled_at":"2026-10-20","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.06,"states_limitation":0.43,"human_participants":0.03,"controlled_comparison":0.92,"randomized_assignment":0.02,"states_quantitative_result":0.99,"compares_with_prior_results":0.97,"mentions_code_or_data_availability":0.98},"evidential_basis":{"p":0.99,"value":"observation or experiment","probabilities":{"argument":0.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.99,"simulation_or_computation":0.01}},"contribution_type":{"p":1.0,"value":"method or tool","probabilities":{"review":0.0,"method_or_tool":1.0,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":0.0}},"contribution_statement":{"p":1.0,"text":"We introduce HarnessPro, a label-free method for evolving harnesses to improve task performance while reducing execution cost.","location":"abstract s2"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":11,"sentences_in_abstract":11},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}