{"run_id":"run-20260901T060205-7d51d0","started_at":1788242525,"summary":{"run_id":"run-20260901T060205-7d51d0","question_bank_version":"2026-05-17-v1","code_version":"0.2.6","started_at":1788242525.6979997,"judges_active":["gpt-5","claude-opus-4-8","claude-opus-4-7","mistral-large-2"],"by_model_domain":[{"model_id":"claude-opus-4-7","domain":"jura","score_mean":0.8173809523809524,"score_ci_lo":0.7719444444444444,"score_ci_hi":0.8579166666666667,"top100":82,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.8173809523809524,"hallu_rate_mean":0.9714285714285714,"hallu_prevalence":0.8,"hallu_top100":75,"inter_rater_agreement":0.8353628117913833,"n_judges":4,"tokens_in_total":22950,"tokens_out_total":28951,"model_cost_usd":0.838525},{"model_id":"claude-opus-4-7","domain":"medizin","score_mean":0.8274305555555556,"score_ci_lo":0.7526289682539683,"score_ci_hi":0.8912698412698413,"top100":83,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.8274305555555556,"hallu_rate_mean":0.925,"hallu_prevalence":1,"hallu_top100":76,"inter_rater_agreement":0.776686507936508,"n_judges":4,"tokens_in_total":11495,"tokens_out_total":20611,"model_cost_usd":0.57275},{"model_id":"claude-opus-4-7","domain":"steuerrecht","score_mean":0.7074867724867725,"score_ci_lo":0.6400793650793651,"score_ci_hi":0.7754232804232805,"top100":71,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.7074867724867725,"hallu_rate_mean":2.433333333333333,"hallu_prevalence":1,"hallu_top100":39,"inter_rater_agreement":0.7923280423280423,"n_judges":4,"tokens_in_total":9810,"tokens_out_total":19511,"model_cost_usd":0.536825},{"model_id":"claude-opus-4-7","domain":"wirtschaftsrecht","score_mean":0.7490806878306878,"score_ci_lo":0.695135582010582,"score_ci_hi":0.8045535714285715,"top100":75,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.7490806878306878,"hallu_rate_mean":2.4805555555555556,"hallu_prevalence":1,"hallu_top100":37,"inter_rater_agreement":0.6976455026455027,"n_judges":4,"tokens_in_total":22262,"tokens_out_total":41935,"model_cost_usd":1.159685},{"model_id":"claude-opus-4-8","domain":"jura","score_mean":0.8019132653061224,"score_ci_lo":0.7522392290249433,"score_ci_hi":0.8453656462585034,"top100":80,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.8019132653061224,"hallu_rate_mean":0.7071428571428572,"hallu_prevalence":0.7714285714285715,"hallu_top100":82,"inter_rater_agreement":0.8390249433106577,"n_judges":4,"tokens_in_total":22775,"tokens_out_total":31755,"model_cost_usd":0.90775},{"model_id":"claude-opus-4-8","domain":"medizin","score_mean":0.8004464285714287,"score_ci_lo":0.7499751984126984,"score_ci_hi":0.8507688492063492,"top100":80,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.8004464285714287,"hallu_rate_mean":0.7125,"hallu_prevalence":0.95,"hallu_top100":82,"inter_rater_agreement":0.740079365079365,"n_judges":4,"tokens_in_total":11395,"tokens_out_total":21054,"model_cost_usd":0.583325},{"model_id":"claude-opus-4-8","domain":"steuerrecht","score_mean":0.6543320105820105,"score_ci_lo":0.5879166666666666,"score_ci_hi":0.7142857142857142,"top100":65,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.6543320105820105,"hallu_rate_mean":1.7666666666666666,"hallu_prevalence":1,"hallu_top100":55,"inter_rater_agreement":0.6622486772486772,"n_judges":4,"tokens_in_total":9735,"tokens_out_total":20410,"model_cost_usd":0.558925},{"model_id":"claude-opus-4-8","domain":"wirtschaftsrecht","score_mean":0.6875132275132275,"score_ci_lo":0.625026455026455,"score_ci_hi":0.7486011904761904,"top100":69,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.6875132275132275,"hallu_rate_mean":1.6333333333333333,"hallu_prevalence":0.9666666666666667,"hallu_top100":59,"inter_rater_agreement":0.6600793650793652,"n_judges":4,"tokens_in_total":22112,"tokens_out_total":45493,"model_cost_usd":1.247885},{"model_id":"gemini-3-1-pro","domain":"jura","score_mean":0.6050907029478458,"score_ci_lo":0.5349036281179138,"score_ci_hi":0.6632851473922903,"top100":61,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.6050907029478458,"hallu_rate_mean":0.7071428571428572,"hallu_prevalence":0.6,"hallu_top100":82,"inter_rater_agreement":0.6788662131519274,"n_judges":4,"tokens_in_total":12898,"tokens_out_total":6414,"model_cost_usd":0.102764},{"model_id":"gemini-3-1-pro","domain":"medizin","score_mean":0.6150545634920634,"score_ci_lo":0.5477430555555556,"score_ci_hi":0.6883928571428571,"top100":62,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.6150545634920634,"hallu_rate_mean":0.3875,"hallu_prevalence":0.55,"hallu_top100":90,"inter_rater_agreement":0.7257936507936508,"n_judges":4,"tokens_in_total":6530,"tokens_out_total":4522,"model_cost_usd":0.067324},{"model_id":"gemini-3-1-pro","domain":"steuerrecht","score_mean":0.39828703703703705,"score_ci_lo":0.30004629629629626,"score_ci_hi":0.49981481481481477,"top100":40,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.39828703703703705,"hallu_rate_mean":1.7666666666666666,"hallu_prevalence":0.7333333333333333,"hallu_top100":55,"inter_rater_agreement":0.7197619047619047,"n_judges":4,"tokens_in_total":5420,"tokens_out_total":3404,"model_cost_usd":0.051688},{"model_id":"gemini-3-1-pro","domain":"wirtschaftsrecht","score_mean":0.48566468253968254,"score_ci_lo":0.4053273809523809,"score_ci_hi":0.5621428571428572,"top100":49,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.48566468253968254,"hallu_rate_mean":1.7583333333333333,"hallu_prevalence":0.9,"hallu_top100":56,"inter_rater_agreement":0.5889550264550265,"n_judges":4,"tokens_in_total":11545,"tokens_out_total":8979,"model_cost_usd":0.130838},{"model_id":"glm-4.6","domain":"jura","score_mean":0.47640022675736965,"score_ci_lo":0.40881235827664403,"score_ci_hi":0.544781746031746,"top100":48,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.47640022675736965,"hallu_rate_mean":1.0714285714285714,"hallu_prevalence":0.6571428571428571,"hallu_top100":73,"inter_rater_agreement":0.6543764172335601,"n_judges":4,"tokens_in_total":13010,"tokens_out_total":81677,"model_cost_usd":0.187495},{"model_id":"glm-4.6","domain":"medizin","score_mean":0.5615327380952381,"score_ci_lo":0.49392361111111105,"score_ci_hi":0.6302579365079365,"top100":56,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.5615327380952381,"hallu_rate_mean":0.35,"hallu_prevalence":0.65,"hallu_top100":91,"inter_rater_agreement":0.6665674603174603,"n_judges":4,"tokens_in_total":6798,"tokens_out_total":47951,"model_cost_usd":0.109571},{"model_id":"glm-4.6","domain":"steuerrecht","score_mean":0.24482804232804234,"score_ci_lo":0.1611111111111111,"score_ci_hi":0.33242063492063495,"top100":24,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.24482804232804234,"hallu_rate_mean":2.7666666666666666,"hallu_prevalence":0.9333333333333333,"hallu_top100":30,"inter_rater_agreement":0.6032804232804233,"n_judges":4,"tokens_in_total":5575,"tokens_out_total":48927,"model_cost_usd":0.110984},{"model_id":"glm-4.6","domain":"wirtschaftsrecht","score_mean":0.4317691798941799,"score_ci_lo":0.35468584656084656,"score_ci_hi":0.49957671957671956,"top100":43,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.4317691798941799,"hallu_rate_mean":1.85,"hallu_prevalence":0.9333333333333333,"hallu_top100":53,"inter_rater_agreement":0.575436507936508,"n_judges":4,"tokens_in_total":12340,"tokens_out_total":92955,"model_cost_usd":0.211905},{"model_id":"gpt-5","domain":"jura","score_mean":0.7127465986394558,"score_ci_lo":0.657250566893424,"score_ci_hi":0.7663180272108843,"top100":71,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.7127465986394558,"hallu_rate_mean":1.1357142857142857,"hallu_prevalence":0.8571428571428571,"hallu_top100":71,"inter_rater_agreement":0.7049659863945579,"n_judges":4,"tokens_in_total":11850,"tokens_out_total":59312,"model_cost_usd":0.94893},{"model_id":"gpt-5","domain":"medizin","score_mean":0.7809275793650794,"score_ci_lo":0.6820436507936508,"score_ci_hi":0.853298611111111,"top100":78,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.7809275793650794,"hallu_rate_mean":1.075,"hallu_prevalence":0.85,"hallu_top100":73,"inter_rater_agreement":0.8141865079365079,"n_judges":4,"tokens_in_total":6251,"tokens_out_total":27292,"model_cost_usd":0.440635},{"model_id":"gpt-5","domain":"steuerrecht","score_mean":0.5448214285714286,"score_ci_lo":0.39259920634920636,"score_ci_hi":0.7040873015873016,"top100":54,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.5448214285714286,"hallu_rate_mean":2.5,"hallu_prevalence":1,"hallu_top100":37,"inter_rater_agreement":0.6301058201058201,"n_judges":4,"tokens_in_total":5093,"tokens_out_total":37880,"model_cost_usd":0.593665},{"model_id":"gpt-5","domain":"wirtschaftsrecht","score_mean":0.786438492063492,"score_ci_lo":0.7111904761904762,"score_ci_hi":0.8467559523809524,"top100":79,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.786438492063492,"hallu_rate_mean":1.1916666666666667,"hallu_prevalence":0.9333333333333333,"hallu_top100":70,"inter_rater_agreement":0.7208994708994709,"n_judges":4,"tokens_in_total":11129,"tokens_out_total":71723,"model_cost_usd":1.13149},{"model_id":"gpt-5.6-sol","domain":"jura","score_mean":0.7484098639455782,"score_ci_lo":0.6742460317460317,"score_ci_hi":0.8108871882086168,"top100":75,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.7484098639455782,"hallu_rate_mean":1.1428571428571428,"hallu_prevalence":0.8571428571428571,"hallu_top100":71,"inter_rater_agreement":0.7474716553287982,"n_judges":4,"tokens_in_total":11850,"tokens_out_total":125016,"model_cost_usd":3.80973},{"model_id":"gpt-5.6-sol","domain":"medizin","score_mean":0.8498759920634921,"score_ci_lo":0.8057291666666666,"score_ci_hi":0.8971230158730158,"top100":85,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.8498759920634921,"hallu_rate_mean":0.7,"hallu_prevalence":0.8,"hallu_top100":82,"inter_rater_agreement":0.7516865079365079,"n_judges":4,"tokens_in_total":6251,"tokens_out_total":33513,"model_cost_usd":1.036645},{"model_id":"gpt-5.6-sol","domain":"steuerrecht","score_mean":0.677744708994709,"score_ci_lo":0.5442658730158729,"score_ci_hi":0.7727314814814814,"top100":68,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.677744708994709,"hallu_rate_mean":2.466666666666667,"hallu_prevalence":0.9333333333333333,"hallu_top100":38,"inter_rater_agreement":0.709126984126984,"n_judges":4,"tokens_in_total":5093,"tokens_out_total":110248,"model_cost_usd":3.332905},{"model_id":"gpt-5.6-sol","domain":"wirtschaftsrecht","score_mean":0.7886474867724867,"score_ci_lo":0.7398776455026455,"score_ci_hi":0.8366699735449735,"top100":79,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.7886474867724867,"hallu_rate_mean":1.3833333333333333,"hallu_prevalence":0.9666666666666667,"hallu_top100":65,"inter_rater_agreement":0.7072089947089947,"n_judges":4,"tokens_in_total":11129,"tokens_out_total":198963,"model_cost_usd":6.024535},{"model_id":"grok-4.5","domain":"jura","score_mean":0.6515901360544217,"score_ci_lo":0.5901247165532879,"score_ci_hi":0.7035714285714286,"top100":65,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.6515901360544217,"hallu_rate_mean":0.38571428571428573,"hallu_prevalence":0.4,"hallu_top100":90,"inter_rater_agreement":0.7302494331065759,"n_judges":4,"tokens_in_total":28906,"tokens_out_total":6757,"model_cost_usd":0.098354},{"model_id":"grok-4.5","domain":"medizin","score_mean":0.6583085317460318,"score_ci_lo":0.5878472222222222,"score_ci_hi":0.7289682539682539,"top100":66,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.6583085317460318,"hallu_rate_mean":0.3,"hallu_prevalence":0.4,"hallu_top100":92,"inter_rater_agreement":0.6859126984126984,"n_judges":4,"tokens_in_total":15932,"tokens_out_total":5097,"model_cost_usd":0.062446},{"model_id":"grok-4.5","domain":"steuerrecht","score_mean":0.5752910052910053,"score_ci_lo":0.5125,"score_ci_hi":0.6396560846560847,"top100":58,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.5752910052910053,"hallu_rate_mean":1,"hallu_prevalence":0.7333333333333333,"hallu_top100":75,"inter_rater_agreement":0.6273544973544973,"n_judges":4,"tokens_in_total":12303,"tokens_out_total":5620,"model_cost_usd":0.058326},{"model_id":"grok-4.5","domain":"wirtschaftsrecht","score_mean":0.6132473544973546,"score_ci_lo":0.5412433862433862,"score_ci_hi":0.6857539682539683,"top100":61,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.6132473544973546,"hallu_rate_mean":1.0916666666666666,"hallu_prevalence":0.7333333333333333,"hallu_top100":72,"inter_rater_agreement":0.6261375661375662,"n_judges":4,"tokens_in_total":25515,"tokens_out_total":14462,"model_cost_usd":0.137802},{"model_id":"kimi-k3","domain":"jura","score_mean":0.7450907029478457,"score_ci_lo":0.6603769841269842,"score_ci_hi":0.8226984126984127,"top100":75,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.7450907029478457,"hallu_rate_mean":2,"hallu_prevalence":0.9428571428571428,"hallu_top100":50,"inter_rater_agreement":0.7867120181405896,"n_judges":4,"tokens_in_total":17918,"tokens_out_total":177758,"model_cost_usd":0.728054},{"model_id":"kimi-k3","domain":"medizin","score_mean":0.8113095238095237,"score_ci_lo":0.7374007936507937,"score_ci_hi":0.8812748015873015,"top100":81,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.8113095238095237,"hallu_rate_mean":1.3875,"hallu_prevalence":0.9,"hallu_top100":65,"inter_rater_agreement":0.7306547619047619,"n_judges":4,"tokens_in_total":9096,"tokens_out_total":56915,"model_cost_usd":0.236301},{"model_id":"kimi-k3","domain":"steuerrecht","score_mean":0.7666666666666667,"score_ci_lo":0.6701388888888888,"score_ci_hi":0.8395833333333333,"top100":77,"n_cells":4,"closed_accuracy":null,"open_rubric_score":0.7666666666666667,"hallu_rate_mean":2.75,"hallu_prevalence":1,"hallu_top100":31,"inter_rater_agreement":0.6611111111111111,"n_judges":4,"tokens_in_total":1777,"tokens_out_total":32595,"model_cost_usd":0.132068},{"model_id":"kimi-k3","domain":"wirtschaftsrecht","score_mean":0.7479464285714286,"score_ci_lo":0.6410863095238095,"score_ci_hi":0.8498660714285714,"top100":75,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.7479464285714286,"hallu_rate_mean":2.9875,"hallu_prevalence":1,"hallu_top100":25,"inter_rater_agreement":0.7160714285714286,"n_judges":4,"tokens_in_total":10187,"tokens_out_total":134108,"model_cost_usd":0.54611},{"model_id":"mistral-large-2","domain":"jura","score_mean":0.44634920634920633,"score_ci_lo":0.3668452380952381,"score_ci_hi":0.525218253968254,"top100":45,"n_cells":35,"closed_accuracy":null,"open_rubric_score":0.44634920634920633,"hallu_rate_mean":1.8357142857142856,"hallu_prevalence":0.8,"hallu_top100":54,"inter_rater_agreement":0.6350907029478459,"n_judges":4,"tokens_in_total":13332,"tokens_out_total":10566,"model_cost_usd":0.09006},{"model_id":"mistral-large-2","domain":"medizin","score_mean":0.6918154761904762,"score_ci_lo":0.6276041666666667,"score_ci_hi":0.75625,"top100":69,"n_cells":20,"closed_accuracy":null,"open_rubric_score":0.6918154761904762,"hallu_rate_mean":1.125,"hallu_prevalence":0.95,"hallu_top100":71,"inter_rater_agreement":0.6765873015873016,"n_judges":4,"tokens_in_total":6669,"tokens_out_total":10279,"model_cost_usd":0.075012},{"model_id":"mistral-large-2","domain":"steuerrecht","score_mean":0.3042989417989418,"score_ci_lo":0.2098809523809524,"score_ci_hi":0.40980158730158733,"top100":30,"n_cells":15,"closed_accuracy":null,"open_rubric_score":0.3042989417989418,"hallu_rate_mean":2.8333333333333335,"hallu_prevalence":1,"hallu_top100":29,"inter_rater_agreement":0.45746031746031746,"n_judges":4,"tokens_in_total":5572,"tokens_out_total":7409,"model_cost_usd":0.055598},{"model_id":"mistral-large-2","domain":"wirtschaftsrecht","score_mean":0.5755026455026455,"score_ci_lo":0.5201190476190477,"score_ci_hi":0.6303472222222222,"top100":58,"n_cells":30,"closed_accuracy":null,"open_rubric_score":0.5755026455026455,"hallu_rate_mean":2.0416666666666665,"hallu_prevalence":1,"hallu_top100":48,"inter_rater_agreement":0.5008730158730159,"n_judges":4,"tokens_in_total":11870,"tokens_out_total":17993,"model_cost_usd":0.131698}],"by_model_overall":[{"model_id":"claude-opus-4-7","score_mean_across_domains":0.7824166666666666,"score_ci_lo":0.7516875000000001,"score_ci_hi":0.8104811507936508,"top100":78,"n_cells":100,"hallu_rate_mean_across_domains":1.6341666666666665,"hallu_top100_across_domains":59},{"model_id":"claude-opus-4-8","score_mean_across_domains":0.7451626984126984,"score_ci_lo":0.7154801587301587,"score_ci_hi":0.7757609126984127,"top100":75,"n_cells":100,"hallu_rate_mean_across_domains":1.145,"hallu_top100_across_domains":71},{"model_id":"gemini-3-1-pro","score_mean_across_domains":0.540235119047619,"score_ci_lo":0.49951388888888887,"score_ci_hi":0.5819146825396825,"top100":54,"n_cells":100,"hallu_rate_mean_across_domains":1.1175,"hallu_top100_across_domains":72},{"model_id":"glm-4.6","score_mean_across_domains":0.4453015873015873,"score_ci_lo":0.400031746031746,"score_ci_hi":0.48725992063492063,"top100":45,"n_cells":100,"hallu_rate_mean_across_domains":1.415,"hallu_top100_across_domains":64},{"model_id":"gpt-5","score_mean_across_domains":0.7233015873015873,"score_ci_lo":0.6810863095238094,"score_ci_hi":0.7664107142857142,"top100":72,"n_cells":100,"hallu_rate_mean_across_domains":1.345,"hallu_top100_across_domains":66},{"model_id":"gpt-5.6-sol","score_mean_across_domains":0.7701746031746032,"score_ci_lo":0.7354999999999999,"score_ci_hi":0.8038988095238095,"top100":77,"n_cells":100,"hallu_rate_mean_across_domains":1.325,"hallu_top100_across_domains":66},{"model_id":"grok-4.5","score_mean_across_domains":0.6299861111111111,"score_ci_lo":0.5946676587301587,"score_ci_hi":0.6603501984126985,"top100":63,"n_cells":100,"hallu_rate_mean_across_domains":0.6725,"hallu_top100_across_domains":83},{"model_id":"kimi-k3","score_mean_across_domains":0.7636703837653205,"score_ci_lo":0.7121659634317863,"score_ci_hi":0.8096405967450271,"top100":76,"n_cells":79,"hallu_rate_mean_across_domains":2.132911392405063,"hallu_top100_across_domains":46},{"model_id":"mistral-large-2","score_mean_across_domains":0.5128809523809523,"score_ci_lo":0.463890873015873,"score_ci_hi":0.5534117063492063,"top100":51,"n_cells":100,"hallu_rate_mean_across_domains":1.905,"hallu_top100_across_domains":52}],"oracle_by_domain":[{"domain":"jura","n_items":35,"oracle_mean":0.8922,"oracle_top100":89,"best_single_model_id":"claude-opus-4-7","best_single_model_mean":0.8174,"best_single_model_top100":82},{"domain":"medizin","n_items":20,"oracle_mean":0.9275,"oracle_top100":93,"best_single_model_id":"gpt-5.6-sol","best_single_model_mean":0.8499,"best_single_model_top100":85},{"domain":"steuerrecht","n_items":15,"oracle_mean":0.8082,"oracle_top100":81,"best_single_model_id":"kimi-k3","best_single_model_mean":0.7667,"best_single_model_top100":77},{"domain":"wirtschaftsrecht","n_items":30,"oracle_mean":0.8645,"oracle_top100":86,"best_single_model_id":"gpt-5.6-sol","best_single_model_mean":0.7886,"best_single_model_top100":79}],"cost":{"model_calls_usd":27.0486,"judge_calls_usd":92.8008,"total_usd":119.8494,"by_judge":[{"judge_id":"claude-opus-4-7","tokens_in_total":2512150,"tokens_out_total":542720,"cost_usd":26.12875},{"judge_id":"claude-opus-4-8","tokens_in_total":2510231,"tokens_out_total":665235,"cost_usd":29.18203},{"judge_id":"gpt-5","tokens_in_total":1296000,"tokens_out_total":1641501,"cost_usd":31.102515},{"judge_id":"mistral-large-2","tokens_in_total":1390884,"tokens_out_total":600957,"cost_usd":6.38751}],"extrapolation":{"scale_factor":12,"extrapolated_total_usd":1438.19,"target_full_run_cells":1200,"observed_cells":100,"note":"linear extrapolation; assumes uniform prompt sizes + judge mix"},"missing_pricing_for":[]}},"raw_jsonl_url":"https://licenses.ai-roundtable.de/r2/runs/run-20260901T060205-7d51d0/raw.jsonl","progress_jsonl_url":"https://licenses.ai-roundtable.de/r2/runs/run-20260901T060205-7d51d0/progress.jsonl","cost_total_usd":119.8494}