Cura 1T leads frontier models on 5 of 6 hardest healthcare benchmarks:</p>\n<ul>\n<li>HealthBench Hard: 36.8 (GPT-5.5: 31.5)</li>\n<li>HealthBench Professional: 66.2 (Claude Fable 5: 66.0)</li>\n<li>MedXpertQA-Text: 60.0 (GPT-5.5: 59.6)</li>\n<li>MedXpertQA-Multimodalt: 72.2 (GPT-5.5: 77.1)</li>\n<li>AgentClinic: 79.6 (Claude Opus 4.8: 79.4)</li>\n<li>MedAgentBench-v2: 94.0 (Claude Opus 4.8: 93.7)</li>\n</ul>\n<p><a href=\"https://cdn-uploads.huggingface.co/production/uploads/632cdea254e2c512c8f95b12/sinoswJp-2Ar-K1BNzm0I.png\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/632cdea254e2c512c8f95b12/sinoswJp-2Ar-K1BNzm0I.png\" alt=\"main_comparison\"></a></p>\n","updatedAt":"2026-07-20T03:52:29.791Z","author":{"_id":"632cdea254e2c512c8f95b12","avatarUrl":"/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg","fullname":"Weiran Yao","name":"weirayao","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":9,"isUserFollowing":false}},"numEdits":1,"identifiedLanguage":{"language":"en","probability":0.5801914930343628},"editors":["weirayao"],"editorAvatarUrls":["/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg"],"reactions":[{"reaction":"🔥","users":["actava-ai","moregpuspls","Emmawxx"],"count":3},{"reaction":"🚀","users":["actava-ai","moregpuspls","Emmawxx"],"count":3},{"reaction":"😎","users":["actava-ai"],"count":1},{"reaction":"👍","users":["Emmawxx"],"count":1}],"isReport":false}},{"id":"6a5d9ba5c028c3a08d69c4f6","author":{"_id":"632cdea254e2c512c8f95b12","avatarUrl":"/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg","fullname":"Weiran Yao","name":"weirayao","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":9,"isUserFollowing":false},"createdAt":"2026-07-20T03:53:09.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"How we trained it: RSI (recursive self-improvement).\n\nEach iteration, a training agent plans a target capability, trains the model, evaluates the graded benchmark trajectories, and data agent synthesizes the next data mixture from the failure modes it finds.\n\nHumans gate every keep-or-revert decision. Reverted rounds stay in the record. One raised headline scores while quietly damaging a held-out subset, so we threw it out. The kept rounds add up: +14.6 on HealthBench Hard, +15.9 on HealthBench Professional, +9.3 on MedAgentBench.\n\nhttps://cdn-uploads.huggingface.co/production/uploads/632cdea254e2c512c8f95b12/BsiaKO-bgZdKp-Gc1koNj.mp4\n","html":"<p>How we trained it: RSI (recursive self-improvement).</p>\n<p>Each iteration, a training agent plans a target capability, trains the model, evaluates the graded benchmark trajectories, and data agent synthesizes the next data mixture from the failure modes it finds.</p>\n<p>Humans gate every keep-or-revert decision. Reverted rounds stay in the record. One raised headline scores while quietly damaging a held-out subset, so we threw it out. The kept rounds add up: +14.6 on HealthBench Hard, +15.9 on HealthBench Professional, +9.3 on MedAgentBench.</p>\n<p><video src=\"https://cdn-uploads.huggingface.co/production/uploads/632cdea254e2c512c8f95b12/BsiaKO-bgZdKp-Gc1koNj.mp4\" controls=\"\" class=\"max-w-full!\"></video></p>\n","updatedAt":"2026-07-20T03:53:09.960Z","author":{"_id":"632cdea254e2c512c8f95b12","avatarUrl":"/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg","fullname":"Weiran Yao","name":"weirayao","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":9,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8417842984199524},"editors":["weirayao"],"editorAvatarUrls":["/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg"],"reactions":[{"reaction":"🤯","users":["actava-ai","moregpuspls","Emmawxx"],"count":3},{"reaction":"🔥","users":["actava-ai","Emmawxx"],"count":2}],"isReport":false}},{"id":"6a5db5bdbeca2a1896f6b018","author":{"_id":"632cdea254e2c512c8f95b12","avatarUrl":"/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg","fullname":"Weiran Yao","name":"weirayao","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":9,"isUserFollowing":false},"createdAt":"2026-07-20T05:44:29.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"@librarian-bot recommend","html":"<p><span class=\"SVELTE_PARTIAL_HYDRATER contents\" data-target=\"UserMention\" data-props=\"{&quot;user&quot;:&quot;librarian-bot&quot;}\"><span class=\"inline-block\"><span class=\"contents\"><a href=\"/librarian-bot\">@<span class=\"underline\">librarian-bot</span></a></span> </span></span> recommend</p>\n","updatedAt":"2026-07-20T05:44:29.779Z","author":{"_id":"632cdea254e2c512c8f95b12","avatarUrl":"/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg","fullname":"Weiran Yao","name":"weirayao","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":9,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7918877601623535},"editors":["weirayao"],"editorAvatarUrls":["/avatars/a6d06cdd75861ae7d589f1343d81a5c5.svg"],"reactions":[],"isReport":false},"replies":[{"id":"6a5dbb8e970d16583af41530","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":376,"isUserFollowing":false},"createdAt":"2026-07-20T06:09:18.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [AutoMedBench: Towards Medical AutoResearch with Agentic AI Models](https://huggingface.co/papers/2606.01961) (2026)\n* [Towards Autonomous and Auditable Medical Imaging Model Development](https://huggingface.co/papers/2607.10522) (2026)\n* [HealthAgentBench: A Unified Benchmark Suite of Realistic Agentic Healthcare Environments for Challenging Frontier AI Agents](https://huggingface.co/papers/2606.31179) (2026)\n* [Are LLMs Ready to Assist Physicians? PhysAssistBench for Interactive Doctor-Patient-EHR Assistance](https://huggingface.co/papers/2606.18613) (2026)\n* [EHR-Complex: Benchmarking Medical Agents for Complex Clinical Reasoning](https://huggingface.co/papers/2606.23301) (2026)\n* [MedGuideX: Internalizing Decision Logic from Executable Guidelines into Large Language Models for Clinical Reasoning](https://huggingface.co/papers/2605.26567) (2026)\n* [CAREAgent: Clinical Agent with Structured Reasoning and Tool-Integrated for Order Generation](https://huggingface.co/papers/2606.01094) (2026)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"<p>This is an automated message from the <a href=\"https://huggingface.co/librarian-bots\">Librarian Bot</a>. I found the following papers similar to this paper. </p>\n<p>The following papers were recommended by the Semantic Scholar API </p>\n<ul>\n<li><a href=\"https://huggingface.co/papers/2606.01961\">AutoMedBench: Towards Medical AutoResearch with Agentic AI Models</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.10522\">Towards Autonomous and Auditable Medical Imaging Model Development</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2606.31179\">HealthAgentBench: A Unified Benchmark Suite of Realistic Agentic Healthcare Environments for Challenging Frontier AI Agents</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2606.18613\">Are LLMs Ready to Assist Physicians? PhysAssistBench for Interactive Doctor-Patient-EHR Assistance</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2606.23301\">EHR-Complex: Benchmarking Medical Agents for Complex Clinical Reasoning</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2605.26567\">MedGuideX: Internalizing Decision Logic from Executable Guidelines into Large Language Models for Clinical Reasoning</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2606.01094\">CAREAgent: Clinical Agent with Structured Reasoning and Tool-Integrated for Order Generation</a> (2026)</li>\n</ul>\n<p> Please give a thumbs up to this comment if you found it helpful!</p>\n<p> If you want recommendations for any Paper on Hugging Face checkout <a href=\"https://huggingface.co/spaces/librarian-bots/recommend_similar_papers\">this</a> Space</p>\n<p> You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: <code>@librarian-bot recommend</code></p>\n","updatedAt":"2026-07-20T06:09:18.592Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":376,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7607210874557495},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false,"parentCommentId":"6a5db5bdbeca2a1896f6b018"}}]}],"primaryEmailConfirmed":false,"paper":{"id":"2607.15314","authors":[{"_id":"6a5d83ee6a69ce099f4d6d91","name":"actAVA AI","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d93","name":"Haolin Chen","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d94","name":"Leon Qi","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d95","name":"Steve Brown","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d96","name":"Deon Metelski","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d97","name":"Tao Xia","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d98","name":"Joonyul Lee","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d99","name":"Qixuan Wang","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d9a","name":"Kevin Riley","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d9b","name":"Frank Wang","hidden":false},{"_id":"6a5d83ee6a69ce099f4d6d9c","name":"Weiran Yao","hidden":false}],"publishedAt":"2026-07-15T00:00:00.000Z","submittedOnDailyAt":"2026-07-20T00:00:00.000Z","title":"Cura 1T: Specialized Model for Agentic Healthcare","submittedOnDailyBy":{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user","name":"taesiri"},"summary":"Healthcare spans high-stakes communication, expert reasoning, and workflow execution, yet specialized LLMs that cover these use cases together remain limited. A healthcare model must handle patient consultation, clinical reasoning over text and images, interactive diagnosis, and electronic health record (EHR) tool use. These capabilities fail in different ways, and a narrow update for one task can degrade another. We present Cura 1T, a healthcare-specialized LLM trained through a human-gated self-evolution loop. In each evolution round, a training agent plans a target capability, trains the model, evaluates benchmark trajectories, and refines the data mixture from observed failures. This data-centered loop improves the model through targeted synthetic and curated examples rather than a single generic medical-data update. Across the healthcare evaluation suite, Cura 1T ranks at or near the top among frontier baselines, while remaining competitive on out-of-domain reasoning and agentic benchmarks.","upvotes":32,"discussionId":"6a5d83ee6a69ce099f4d6d9d","projectPage":"https://actava.ai/cura","githubRepo":"https://github.com/actava-ai/Cura","githubRepoAddedBy":"user","githubStars":13},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},{"_id":"68edc310a4f606a8123967e7","avatarUrl":"/avatars/908302a1d06a5e0984af74f483b7ed69.svg","isPro":false,"fullname":"Weiran Yao","user":"weiran-actava","type":"user"},{"_id":"661573234c2f29635e93bb71","avatarUrl":"/avatars/fba95e382454485766b6349d6281b715.svg","isPro":false,"fullname":"Weiran Yao","user":"weiranyao","type":"user"},{"_id":"68edc7f6a17c9d2e3dba81b2","avatarUrl":"/avatars/e0cc4e300e9a2875f866ab04bfc7d4b9.svg","isPro":false,"fullname":"ggpd","user":"moregpuspls","type":"user"},{"_id":"6a14ed379224246011a04d87","avatarUrl":"/avatars/e3d1d655338bf264092bb1101e65f896.svg","isPro":false,"fullname":"Tao Xiong","user":"taoxiong-wch","type":"user"},{"_id":"6a14edbef05031c131840529","avatarUrl":"/avatars/862eae50cafb191b80b26ff8397aa3b7.svg","isPro":false,"fullname":"Yi Yang","user":"dryiyang","type":"user"},{"_id":"65489653cbe50f378d94f14a","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65489653cbe50f378d94f14a/Bhl3yRJ9lYCc7HrYj9jJf.png","isPro":false,"fullname":"Leon Qi","user":"leonqi","type":"user"},{"_id":"64bb3e754b4ff0d509141938","avatarUrl":"/avatars/81617fd676a3d3133811348055e68fc7.svg","isPro":false,"fullname":"Zhepeng Cen","user":"czp16","type":"user"},{"_id":"6a0c6ae1d575c4c20ae1bae9","avatarUrl":"/avatars/27fee9022ef159138de77c1113f620cf.svg","isPro":false,"fullname":"Zhiwei Liu","user":"jimliu96","type":"user"},{"_id":"6553cd2296c5902fa386f57f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6553cd2296c5902fa386f57f/AEDbKbifdmDW4eJf6pUqw.png","isPro":false,"fullname":"Zeyu Tang","user":"zeyutang","type":"user"},{"_id":"5f12485c0c833276f61f1afb","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1595033594228-noauth.jpeg","isPro":false,"fullname":"Xiangchen Song","user":"xiangchensong","type":"user"},{"_id":"64c49396bf1954890192bfbf","avatarUrl":"/avatars/b2dbd1d601911440788efcea2a2b77d3.svg","isPro":false,"fullname":"Yutong Dai","user":"UncleFish","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2607/2607.15314.md","query":{}}">