{"schema_version":2,"configurations":[{"id":"hermes-luna-high-builtin-20260916","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"gpt-5.6-luna-high","name":"GPT 5.6 Luna","provider":"OpenAI"},"memory":{"id":"builtin","name":"Built-in"},"personas":{"morgan":{"total":200,"passes":131,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-builtin-results.json.gz/","sha256":"2fcb137075ddb29a72b14d1eacdb7f013202379fdae37f62b9999ac04246179e"}},"alex":{"total":200,"passes":111,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-builtin-results.json.gz/","sha256":"308e7131053e76a1d56eda23511c59913db6491f0e5e57e0f479ad9f12dc77fd"}},"riley":{"total":200,"passes":152,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-builtin-results.json.gz/","sha256":"0b8b8196f9f9a2eb538676f482b2239930a68189fd05abb010d692772700a218"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/builtin-configuration.zip","sha256":"5faa3c3b70a0670a5388028d094b3e263af0a628d45b2e9d8996b7542221007c"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/builtin-sources.json/","sha256":"561237cdb3e244ad64c17bdf6a00c82d75dac6678c322fd7635958b90840a484"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/builtin-results-and-traces.tar/","sha256":"af8039aa2fd9e49e0444d1c1ff45428a9acbdc0af558e9a295e9dd3f699ecd60"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/builtin-results-and-traces.tar/","sha256":"af8039aa2fd9e49e0444d1c1ff45428a9acbdc0af558e9a295e9dd3f699ecd60"}},"total_cost_usd_test_calls":8.04923834,"median_latency_seconds":44.35470588399994,"p95_latency_seconds":100.88306797600004,"agent_inference_cost_scope":"Token-priced agent inference for all 600 tests; excludes ingestion, grading, and memory-provider service fees.","total":600,"passes":394,"pass_rate":0.6566666666666666,"total_cost_usd":61.482132030000024,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/gpt-5.6-luna/builtin"},{"id":"hermes-luna-high-mem0-20260916","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"gpt-5.6-luna-high","name":"GPT 5.6 Luna","provider":"OpenAI"},"memory":{"id":"mem0","name":"Mem0"},"personas":{"morgan":{"total":200,"passes":156,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-mem0-results.json.gz/","sha256":"38ecb0f0be0619e344fd0f504392ec2c0a6bd8e186248f36d2b319c872c40161"}},"alex":{"total":200,"passes":132,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-mem0-results.json.gz/","sha256":"33476ff28a07ded2e8f205c0b74f26d0c2934d7e5254f62e82bd05b49dcbc86e"}},"riley":{"total":200,"passes":136,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-mem0-results.json.gz/","sha256":"c6397452ee46f383ffe2b0b4e7045fb5b28e14b9499730d7242f3849679d9800"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/mem0-configuration.zip","sha256":"ee75fb1e4fafde02370e45554048a4463c11ed1aff2c01e823054838c09d3e3f"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/mem0-sources.json/","sha256":"ca1068d117af4a607d47559b89d1d1ec7536cd3c1f2b728484cfde03a1cc3f1e"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/mem0-results-and-traces.tar/","sha256":"4bd145825fcaf13ae1080c955b32916f4bf12848d173a032092b3800528dd2ae"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/mem0-results-and-traces.tar/","sha256":"4bd145825fcaf13ae1080c955b32916f4bf12848d173a032092b3800528dd2ae"}},"total_cost_usd_test_calls":3.22984434,"median_latency_seconds":37.68894314099998,"p95_latency_seconds":81.16765088794976,"agent_inference_cost_scope":"Token-priced agent inference for all 600 tests; excludes ingestion, grading, and memory-provider service fees.","total":600,"passes":424,"pass_rate":0.7066666666666667,"total_cost_usd":96.21388145000004,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/gpt-5.6-luna/mem0"},{"id":"hermes-luna-high-honcho-20260916","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"gpt-5.6-luna-high","name":"GPT 5.6 Luna","provider":"OpenAI"},"memory":{"id":"honcho","name":"Honcho"},"personas":{"morgan":{"total":200,"passes":144,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-honcho-results.json.gz/","sha256":"ee9c163a527897166fa2483d133bb92db240dbd976703d868f5fbbc5341a8122"}},"alex":{"total":200,"passes":120,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-honcho-results.json.gz/","sha256":"d1ede67c21c32aee44b353b0e868999fceb41870121bc0ba9fb76b81689ff626"}},"riley":{"total":200,"passes":147,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-honcho-results.json.gz/","sha256":"b4087638890b02c6707fc0b5999dfa36142ed6c03524972372cca48216596832"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/honcho-configuration.zip","sha256":"547626d2b94d00035b1e3d1a19a215146f50354f67589aa6f42cd3b3a650a3f1"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/honcho-sources.json/","sha256":"04461341965fa9158a60334d4834bc2a6e28f1397161770c6ce1bdf52891be15"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/honcho-results-and-traces.tar/","sha256":"95b35ce778c9e05148b3edb93c4c2364a171603ea97ce8520a3fad2d2132ec02"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/honcho-results-and-traces.tar/","sha256":"95b35ce778c9e05148b3edb93c4c2364a171603ea97ce8520a3fad2d2132ec02"}},"total_cost_usd_test_calls":4.93670969,"median_latency_seconds":44.66409763349981,"p95_latency_seconds":94.57841570574976,"agent_inference_cost_scope":"Token-priced agent inference for all 600 tests; excludes ingestion, grading, and memory-provider service fees.","total":600,"passes":411,"pass_rate":0.685,"total_cost_usd":142.85233241179964,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/gpt-5.6-luna/honcho"},{"id":"hermes-luna-high-hindsight-20260916","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"gpt-5.6-luna-high","name":"GPT 5.6 Luna","provider":"OpenAI"},"memory":{"id":"hindsight","name":"Hindsight"},"personas":{"morgan":{"total":200,"passes":144,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-hindsight-results.json.gz/","sha256":"e9402d8c178761ef08180f4e8523cc95367d9121ced26a764db348dc1cd714b8"}},"alex":{"total":200,"passes":127,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-hindsight-results.json.gz/","sha256":"36802caf541dcc6ae159fbbdf2e796e9835119e08b97514c473259421b2724fb"}},"riley":{"total":200,"passes":146,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-hindsight-results.json.gz/","sha256":"5fc0d117f1675c71eb80dcfcf51310d8128259c2a4776825760400fd7d0cc4a2"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/hindsight-configuration.zip","sha256":"eb1c069251c60ecb8e290f4b237eccf48addfa0cf635365de469ed01c0a86b24"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/hindsight-sources.json/","sha256":"43e9a1b4f4c45185aea807e1aa1b780657b2ff5e2e1efda50b118ce21324f7e2"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/hindsight-results-and-traces.tar/","sha256":"3383c53e35a65f4f6dbcd93b917bd5037a16bba6ef414c8828ad1c22791860d2"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/hindsight-results-and-traces.tar/","sha256":"3383c53e35a65f4f6dbcd93b917bd5037a16bba6ef414c8828ad1c22791860d2"}},"total_cost_usd_test_calls":3.84165259,"median_latency_seconds":55.307897638500435,"p95_latency_seconds":158.87933509020013,"agent_inference_cost_scope":"Token-priced agent inference for all 600 tests; excludes ingestion, grading, and memory-provider service fees.","total":600,"passes":417,"pass_rate":0.695,"total_cost_usd":84.64996336682626,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/gpt-5.6-luna/hindsight"},{"id":"hermes-minimax-m3-high-builtin-20260921","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"minimax-m3-high","name":"MiniMax M3","provider":"MiniMax"},"memory":{"id":"builtin","name":"Built-in"},"personas":{"morgan":{"total":200,"passes":56,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-minimax-builtin-results.json.gz/","sha256":"c0ef9559b83b778ce8d6ece2d0e6ac66215780be1a047c35c062f9ba8e5e22b9"}},"alex":{"total":200,"passes":45,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-minimax-builtin-results.json.gz/","sha256":"1b4aed29516e93436b96e3595304c48886a69acb4401780186fe3eeb0a1989b0"}},"riley":{"total":200,"passes":58,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-minimax-builtin-results.json.gz/","sha256":"0f2157ff29358050de0b4aca506aa0cd2c1944dd92bd518929b7019f1cdc0b58"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/minimax-builtin-configuration.zip","sha256":"c873642feac7cdc3b5aff7e51b3c046dd382f6826b3cb38371507ce575a814ed"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-builtin-sources.json/","sha256":"c7d1c001122d179514f15b8cb39b476b5f95b642761ac7352509cd9e4872eb42"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-builtin-results-and-traces.tar.gz/","sha256":"390b9d0a820a05c0f78eea7a6c477a4ad0eb782a685290afc60cee8f5aa30cf3"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-builtin-results-and-traces.tar.gz/","sha256":"390b9d0a820a05c0f78eea7a6c477a4ad0eb782a685290afc60cee8f5aa30cf3"}},"total_cost_usd_test_calls":9.15751068,"median_latency_seconds":51.395184124499906,"p95_latency_seconds":172.2558978655,"agent_inference_cost_scope":"Token-priced agent inference for all 600 tests; excludes ingestion, grading, and other provider fees.","total":600,"passes":159,"pass_rate":0.265,"total_cost_usd":107.03188422,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/minimax-m3/builtin"},{"id":"hermes-minimax-m3-high-mem0-20260920","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"minimax-m3-high","name":"MiniMax M3","provider":"MiniMax"},"memory":{"id":"mem0","name":"Mem0"},"personas":{"morgan":{"total":200,"passes":108,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-minimax-mem0-results.json.gz/","sha256":"88a646c882cf25438a6322d4a0787df0a1a12648137e709a4cf120b3218628ce"}},"alex":{"total":200,"passes":82,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-minimax-mem0-results.json.gz/","sha256":"41f2a237b061b27735053e3320ab63b84ac777645428bb9fc3c16f8bdc611b36"}},"riley":{"total":200,"passes":97,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-minimax-mem0-results.json.gz/","sha256":"b905096299c97c9a182479d6c668b607988fda636a24e52007d1501dde10ec4b"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/minimax-mem0-configuration.zip","sha256":"cf17d2fee60a8606faa3ea2db2e9fc2f88840eba8da4accbd8f983b0f602b641"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-mem0-sources.json/","sha256":"d5c86791256c94c2fd1fe89c1e3017de44bb63460216068047b7643bd5428572"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-mem0-results-and-traces.tar/","sha256":"7da0014e033363fa561a2abd2bb3608caf8256fea0211e904eea4192879ab695"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-mem0-results-and-traces.tar/","sha256":"7da0014e033363fa561a2abd2bb3608caf8256fea0211e904eea4192879ab695"}},"total_cost_usd_test_calls":5.02433952,"median_latency_seconds":46.95490127649964,"p95_latency_seconds":266.30208848680013,"agent_inference_cost_scope":"Token-priced agent inference for all 600 tests; excludes ingestion, grading, and other provider fees.","total":600,"passes":287,"pass_rate":0.47833333333333333,"total_cost_usd":148.0180480712263,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/minimax-m3/mem0"},{"id":"hermes-gpt-5.6-luna-high-supermemory-20260921","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"gpt-5.6-luna-high","name":"GPT 5.6 Luna","provider":"OpenAI"},"memory":{"id":"supermemory","name":"Supermemory"},"personas":{"morgan":{"total":200,"passes":138,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-luna-supermemory-results.json.gz/","sha256":"203a84b383a4ada9a51130a51a166832021218ecd3e7e6df39f75021b60ea961"}},"alex":{"total":200,"passes":98,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-luna-supermemory-results.json.gz/","sha256":"b8a4c53ed4d6b64068d773e1e08cc84f21b379b750959ab68ab1d0e08b8f2a7e"}},"riley":{"total":200,"passes":119,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-luna-supermemory-results.json.gz/","sha256":"8b0fb0c62cd34f59172ad465451dd7330f78bd77b7966f217dbb6c2444b0c7a0"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/luna-supermemory-configuration.zip","sha256":"01dc011d029e1111b0b50eb1424e072bb53c238ac41ec103ea6d0384f9d1bea9"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/luna-supermemory-sources.json/","sha256":"c97c531f4f61501b60cc9a8007231882bab506fd6bdacb7bc30ebf557273ace2"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/luna-supermemory-results-and-traces.tar.gz/","sha256":"2d4a57e7dff4fe51f939f6c1839697f05df1fb7563d2076c9617cc68598c4499"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/luna-supermemory-results-and-traces.tar.gz/","sha256":"2d4a57e7dff4fe51f939f6c1839697f05df1fb7563d2076c9617cc68598c4499"}},"total_cost_usd_test_calls":7.38471046,"median_latency_seconds":52.67815892099995,"p95_latency_seconds":134.7041816865998,"agent_inference_cost_scope":"Agent inference during evaluation; excludes ingestion, grading, memory processing, and infrastructure.","total_cost_usd":345.6447958346001,"total":600,"passes":355,"pass_rate":0.5916666666666667,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/gpt-5.6-luna/supermemory"},{"id":"hermes-minimax-m3-high-honcho-20260921","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"minimax-m3-high","name":"MiniMax M3","provider":"MiniMax"},"memory":{"id":"honcho","name":"Honcho"},"personas":{"morgan":{"total":200,"passes":108,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-minimax-honcho-results.json.gz/","sha256":"e12520ef244d2e9d9a2e7fc7e12adbb3a07127f9e5da7605338902064d18f59b"}},"alex":{"total":200,"passes":91,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-minimax-honcho-results.json.gz/","sha256":"045cd6cc0c76736bcd811c4b9486bcf066896e3e73c08d2d2d793eeb4f599be0"}},"riley":{"total":200,"passes":85,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-minimax-honcho-results.json.gz/","sha256":"3b104493fc2bb774c39df10bfca4a7edfe8f3a73d88fc12743f8777da44cc014"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/minimax-honcho-configuration.zip","sha256":"6cabf7e4c5d481e0404ae1dd02a7dc68d52d0625c3219d4dd44ec1033abd87da"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-honcho-sources.json/","sha256":"572441b005da6467388442061a6374798164ed4fa2cb0354c566470defaaac58"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-honcho-results-and-traces.tar.gz/","sha256":"f743447b642c8bf5a979cf7a925d765fb7775f04d1736298837d9d30b85703ec"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-honcho-results-and-traces.tar.gz/","sha256":"f743447b642c8bf5a979cf7a925d765fb7775f04d1736298837d9d30b85703ec"}},"total_cost_usd_test_calls":7.040642579999999,"median_latency_seconds":80.32813606000036,"p95_latency_seconds":318.56076525949953,"agent_inference_cost_scope":"Agent inference during evaluation; excludes ingestion, grading, memory processing, and infrastructure.","total_cost_usd":211.84779666729017,"total":600,"passes":284,"pass_rate":0.4733333333333334,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/minimax-m3/honcho"},{"id":"hermes-minimax-m3-high-hindsight-20260921","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"minimax-m3-high","name":"MiniMax M3","provider":"MiniMax"},"memory":{"id":"hindsight","name":"Hindsight"},"personas":{"morgan":{"total":200,"passes":109,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-minimax-hindsight-results.json.gz/","sha256":"a8d43a4b4db3d90e5157cc6ec5a26d4ed1d0438c94b273daa69ba036d0071b5a"}},"alex":{"total":200,"passes":93,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-minimax-hindsight-results.json.gz/","sha256":"1c50a2f8b7f34ea55d998cf28b3ccd9c5d732a4e761bbf8f1d126d5d98ff3fd4"}},"riley":{"total":200,"passes":44,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-minimax-hindsight-results.json.gz/","sha256":"05a1bc96a72991824ade99bf158d1fdb2f1b9dfa056b1bf592d1a0c3e513ffe4"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/minimax-hindsight-configuration.zip","sha256":"77d362659f54e3aba20c37666801baedc2d61917dcc01862dc00a1ed5a4c7595"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-hindsight-sources.json/","sha256":"1d49f52992ba3026bbe282e269a4ec13aad2895b3d2a92ed8cf7621eee4bc301"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-hindsight-results-and-traces.tar.gz/","sha256":"9b99ca95f3c76a488d37f23446829f09e777f28ee7562743fe7e6cfd219bca29"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-hindsight-results-and-traces.tar.gz/","sha256":"9b99ca95f3c76a488d37f23446829f09e777f28ee7562743fe7e6cfd219bca29"}},"total_cost_usd_test_calls":9.97001916,"median_latency_seconds":144.27637541699983,"p95_latency_seconds":602.8579876240501,"agent_inference_cost_scope":"Agent inference during evaluation; excludes ingestion, grading, memory processing, and infrastructure.","total_cost_usd":158.3259366776085,"total":600,"passes":246,"pass_rate":0.41,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/minimax-m3/hindsight"},{"id":"hermes-minimax-m3-high-supermemory-20260921","harness":{"id":"hermes","name":"Hermes"},"model":{"id":"minimax-m3-high","name":"MiniMax M3","provider":"MiniMax"},"memory":{"id":"supermemory","name":"Supermemory"},"personas":{"morgan":{"total":200,"passes":45,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-minimax-supermemory-results.json.gz/","sha256":"8b91022d18b835b641ecfda8d20676dd422a2b1f43cc6d9fd9ee6432a7decc86"}},"alex":{"total":200,"passes":37,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-minimax-supermemory-results.json.gz/","sha256":"c5348806fbe88b8d605994b55398d3dc3459428bda98c2dd4faff392f3e0d6eb"}},"riley":{"total":200,"passes":38,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-minimax-supermemory-results.json.gz/","sha256":"56ca7ad7033b24c523a9aa02d69e258f14091ecfcff7a5ac90cd88a48fbba0a5"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/minimax-supermemory-configuration.zip","sha256":"df7f1b8b4a33fbd5c70b56940912e16f68b2bd9fe898bdf2ec1c04baee990f41"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-supermemory-sources.json/","sha256":"ec247efdf3d9b1ff84acf7200c71238fe452f6f6b033a8d594442e36d89ac025"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-supermemory-results-and-traces.tar.gz/","sha256":"ab0c02a97e0cc1d3eff776787242b2151615a14b12a6d1e8d793cd7f0066f7af"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/minimax-supermemory-results-and-traces.tar.gz/","sha256":"ab0c02a97e0cc1d3eff776787242b2151615a14b12a6d1e8d793cd7f0066f7af"}},"total_cost_usd_test_calls":6.622469639999999,"median_latency_seconds":137.52324065899964,"p95_latency_seconds":1846.17102980525,"agent_inference_cost_scope":"Agent inference during evaluation; excludes ingestion, grading, memory processing, and infrastructure.","total_cost_usd":497.7013825211592,"total":600,"passes":120,"pass_rate":0.2,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/hermes/minimax-m3/supermemory"},{"id":"claude-code-sonnet-5-builtin-20260921","harness":{"id":"claude-code","name":"Claude Code"},"model":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","provider":"Anthropic"},"memory":{"id":"builtin","name":"Built-in"},"personas":{"morgan":{"total":200,"passes":69,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-claude-builtin-results.json.gz/","sha256":"c6e091e4a47d31fb94c3d2b73ce392cca62bf43149ea124e84c61ad4ad06a5d3"}},"alex":{"total":200,"passes":47,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-claude-builtin-results.json.gz/","sha256":"b3474a604c3a1bcbb55d5eb7c33de9fc948866401c8e3c6f0a59f5dd0c8572a0"}},"riley":{"total":200,"passes":42,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-claude-builtin-results.json.gz/","sha256":"f17c317a5be8738090e3ca52eed5696274342e5739a3d8cc35e5d996a8fb3567"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/claude-builtin-configuration.zip","sha256":"b7faf2a5ff617654d0b171024c94a13e206eb4e695ef6126f91b213553c6cb34"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-builtin-sources.json/","sha256":"4cd0bd98db520722ac3e3dcf48cfdc1a6976292c7d1dcf0c1bde249a5cf619f2"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-builtin-results-and-traces.tar.gz/","sha256":"8f8c04305aee0fc101f78e2228b6f64e8747d2f50cee3276f1b3882035d3b464"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-builtin-results-and-traces.tar.gz/","sha256":"8f8c04305aee0fc101f78e2228b6f64e8747d2f50cee3276f1b3882035d3b464"}},"total_cost_usd_test_calls":87.8874987,"median_latency_seconds":32.19062741699963,"p95_latency_seconds":68.7752687529997,"agent_inference_cost_scope":"Token-priced Claude agent inference for all 600 tests; excludes ingestion, grading, and other fees.","total_cost_usd":1132.74661225,"total":600,"passes":158,"pass_rate":0.2633333333333333,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/claude-code/claude-sonnet-5/builtin"},{"id":"claude-code-claude-sonnet-5-honcho-20260921","harness":{"id":"claude-code","name":"Claude Code"},"model":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","provider":"Anthropic"},"memory":{"id":"honcho","name":"Honcho"},"personas":{"morgan":{"total":200,"passes":93,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-claude-honcho-results.json.gz/","sha256":"d4c406ab14276c8ea63136040daa5b55191576ed24cce4315887d9aaf3d757d0"}},"alex":{"total":200,"passes":63,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-claude-honcho-results.json.gz/","sha256":"5b6fc4ca5e20698ec211282c8ed759139c55487e55c35f3bead0bedcab9ce14a"}},"riley":{"total":200,"passes":59,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-claude-honcho-results.json.gz/","sha256":"1996ac3a4fa02392c41f24b340f3e580df3ca5dc65d4efdc5e0cdabc68923b35"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/claude-honcho-configuration.zip","sha256":"411b513f61153deba128b01949caf6f36d413a07941bdeda8965b55ce72949d8"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-honcho-sources.json/","sha256":"08d932d8f5e4fff5e01525d5724f9532724e1bfe68e5c693c438bd121a492c3e"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-honcho-results-and-traces.tar.gz/","sha256":"250cf1a17cb32fdde1bbf1a025d72aef9cd15b85ab3eb176e8c04f8509abd70a"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-honcho-results-and-traces.tar.gz/","sha256":"250cf1a17cb32fdde1bbf1a025d72aef9cd15b85ab3eb176e8c04f8509abd70a"}},"total_cost_usd_test_calls":95.7829394,"median_latency_seconds":37.923294940000005,"p95_latency_seconds":78.54438862180032,"agent_inference_cost_scope":"Agent inference during evaluation; excludes ingestion, grading, memory processing, and infrastructure.","total_cost_usd":1565.8183545579857,"total":600,"passes":215,"pass_rate":0.35833333333333334,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/claude-code/claude-sonnet-5/honcho"},{"id":"claude-code-claude-sonnet-5-mem0-20260921","harness":{"id":"claude-code","name":"Claude Code"},"model":{"id":"claude-sonnet-5","name":"Claude Sonnet 5","provider":"Anthropic"},"memory":{"id":"mem0","name":"Mem0"},"personas":{"morgan":{"total":200,"passes":81,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/morgan-claude-mem0-results.json.gz/","sha256":"385c129e3996da8c0aa5558c41ab6a24798a6557c4d68358aa3f32adba727ba5"}},"alex":{"total":200,"passes":72,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/alex-claude-mem0-results.json.gz/","sha256":"5452e053ed4661e655506e09b6872593148ac6be1ebedf50e5b13721d8809af3"}},"riley":{"total":200,"passes":41,"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/riley-claude-mem0-results.json.gz/","sha256":"65b923dee5b51ea276ae9ff7400373649d86b9dd88b3ac91acaf0e08e6d4b549"}}},"evidence":{"configuration":{"url":"https://dolphinbench.ai/leaderboard/cost-records/claude-mem0-configuration.zip","sha256":"0afde104822a121d32091b1ec64dc995d2373be98a4b734458e09e16e15c972b"},"source":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-mem0-sources.json/","sha256":"41abec2ff6d8881c01d56bfca89c3ef71b43661da8ce88a9367e8cf3b51b4018"},"grades":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-mem0-results-and-traces.tar.gz/","sha256":"839e33975dce834b0382677745d345144f71804e4da4cfc573ae7c85c96d831e"},"recordings":{"url":"https://dolphinbench-review.vercel.app/leaderboard/evidence/claude-mem0-results-and-traces.tar.gz/","sha256":"839e33975dce834b0382677745d345144f71804e4da4cfc573ae7c85c96d831e"}},"total_cost_usd_test_calls":107.4124817,"median_latency_seconds":42.29555081899973,"p95_latency_seconds":153.8408404088997,"agent_inference_cost_scope":"Agent inference during evaluation; excludes ingestion, grading, memory processing, and infrastructure.","total_cost_usd":1830.56848095,"total":600,"passes":194,"pass_rate":0.3233333333333333,"total_cost_scope":"Cost includes agent and memory-system processing during ingestion and testing.","runs_url":"https://github.com/mem0ai/dolphinbench/tree/main/results/claude-code/claude-sonnet-5/mem0"}],"release_sha256":"40107f293243a6cb3941c16e75ce2e3495ed31c68a1015b8537efeb8dfec2ae9"}