Librarian Bot. I found the following papers similar to this paper.

\n

The following papers were recommended by the Semantic Scholar API

\n

\n
• Apriel-Nemotron-15B-Thinker (2025)
\n
• InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency (2025)
\n
• K2-Think: A Parameter-Efficient Reasoning System (2025)
\n
• MobileLLM-R1: Exploring the Limits of Sub-Billion Language Model Reasoners with Open Training Recipes (2025)
\n
• VARCO-VISION-2.0 Technical Report (2025)
\n
• SAIL-VL2 Technical Report (2025)
\n
• VLA-R1: Enhancing Reasoning in Vision-Language-Action Models (2025)
\n

\n

Please give a thumbs up to this comment if you found it helpful!

\n

If you want recommendations for any Paper on Hugging Face checkout this Space

\n

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2025-10-07T01:34:51.212Z","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":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.687838077545166},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}},{"id":"68e50c9fa5dc0051d1be4152","author":{"_id":"64a16b1aeacb4b50ba1c889d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64a16b1aeacb4b50ba1c889d/EhWoBGL6LlFGIVTUsp2F4.jpeg","fullname":"Lai Wei","name":"WaltonFuture","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10,"isUserFollowing":false},"createdAt":"2025-10-07T12:50:39.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"Thanks for your work! I have a quick question: how do you organize the data formats for tasks like Image Reconstruction and Visual Matching in CPT Stage 2? I think this synthetic augmentation approach is particularly interesting.\n\nThank you!","html":"

Thanks for your work! I have a quick question: how do you organize the data formats for tasks like Image Reconstruction and Visual Matching in CPT Stage 2? I think this synthetic augmentation approach is particularly interesting.

\n

Thank you!

\n","updatedAt":"2025-10-07T12:50:39.525Z","author":{"_id":"64a16b1aeacb4b50ba1c889d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64a16b1aeacb4b50ba1c889d/EhWoBGL6LlFGIVTUsp2F4.jpeg","fullname":"Lai Wei","name":"WaltonFuture","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":10,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9234086871147156},"editors":["WaltonFuture"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/64a16b1aeacb4b50ba1c889d/EhWoBGL6LlFGIVTUsp2F4.jpeg"],"reactions":[{"reaction":"👍","users":["victor"],"count":1}],"isReport":false}},{"id":"68f401812d5098be2eb37450","author":{"_id":"65d9fc2a0e6ad24551d87a1e","avatarUrl":"/avatars/3aedb9522cc3cd08349d654f523fd792.svg","fullname":"Grant Singleton","name":"grantsing","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":4,"isUserFollowing":false},"createdAt":"2025-10-18T21:07:13.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"arXiv explained breakdown of this paper 👉 https://arxivexplained.com/papers/apriel-15-15b-thinker","html":"

arXiv explained breakdown of this paper 👉 https://arxivexplained.com/papers/apriel-15-15b-thinker

\n","updatedAt":"2025-10-18T21:07:13.741Z","author":{"_id":"65d9fc2a0e6ad24551d87a1e","avatarUrl":"/avatars/3aedb9522cc3cd08349d654f523fd792.svg","fullname":"Grant Singleton","name":"grantsing","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":4,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7057124376296997},"editors":["grantsing"],"editorAvatarUrls":["/avatars/3aedb9522cc3cd08349d654f523fd792.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2510.01141","authors":[{"_id":"68de25d26024653e8a3ed204","name":"Shruthan Radhakrishna","hidden":false},{"_id":"68de25d26024653e8a3ed205","user":{"_id":"62d913739a5353eef9d7edf3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d913739a5353eef9d7edf3/pRgen2izGJle3ahupOdC7.jpeg","isPro":false,"fullname":"Aman Tiwari","user":"amant555","type":"user"},"name":"Aman Tiwari","status":"claimed_verified","statusLastChangedAt":"2025-10-02T13:54:38.519Z","hidden":false},{"_id":"68de25d26024653e8a3ed206","name":"Aanjaneya Shukla","hidden":false},{"_id":"68de25d26024653e8a3ed207","name":"Masoud Hashemi","hidden":false},{"_id":"68de25d26024653e8a3ed208","name":"Rishabh Maheshwary","hidden":false},{"_id":"68de25d26024653e8a3ed209","name":"Shiva Krishna Reddy Malay","hidden":false},{"_id":"68de25d26024653e8a3ed20a","name":"Jash Mehta","hidden":false},{"_id":"68de25d26024653e8a3ed20b","name":"Pulkit Pattnaik","hidden":false},{"_id":"68de25d26024653e8a3ed20c","name":"Saloni Mittal","hidden":false},{"_id":"68de25d26024653e8a3ed20d","name":"Khalil Slimi","hidden":false},{"_id":"68de25d26024653e8a3ed20e","name":"Kelechi Ogueji","hidden":false},{"_id":"68de25d26024653e8a3ed20f","user":{"_id":"60c8e5a203ddd410286bad87","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1623778630501-noauth.jpeg","isPro":false,"fullname":"Akintunde Oladipo","user":"theyorubayesian","type":"user"},"name":"Akintunde Oladipo","status":"claimed_verified","statusLastChangedAt":"2025-10-13T10:10:36.940Z","hidden":false},{"_id":"68de25d26024653e8a3ed210","name":"Soham Parikh","hidden":false},{"_id":"68de25d26024653e8a3ed211","user":{"_id":"66845714d4e3eff8e18d8dda","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/-LV096QGIbXnGGjjCad4a.jpeg","isPro":false,"fullname":"Nifemi Bamgbose","user":"onifemibam","type":"user"},"name":"Oluwanifemi Bamgbose","status":"claimed_verified","statusLastChangedAt":"2025-10-22T15:47:00.117Z","hidden":false},{"_id":"68de25d26024653e8a3ed212","name":"Toby Liang","hidden":false},{"_id":"68de25d26024653e8a3ed213","name":"Ahmed Masry","hidden":false},{"_id":"68de25d26024653e8a3ed214","name":"Khyati Mahajan","hidden":false},{"_id":"68de25d26024653e8a3ed215","user":{"_id":"642f99079b2484d7d857341b","avatarUrl":"/avatars/01965cc5a5dbe9c08025a51973462a6a.svg","isPro":false,"fullname":"Sai Rajeswar","user":"rajeswarsai","type":"user"},"name":"Sai Rajeswar Mudumba","status":"claimed_verified","statusLastChangedAt":"2025-10-05T12:47:29.534Z","hidden":false},{"_id":"68de25d26024653e8a3ed216","name":"Vikas Yadav","hidden":false},{"_id":"68de25d26024653e8a3ed217","user":{"_id":"63d3095c2727d7888cbb54e2","avatarUrl":"/avatars/3d437afdf19cb43c6b67a15e4c2955f8.svg","isPro":false,"fullname":"Sathwik Tejaswi Madhusudhan","user":"stm4","type":"user"},"name":"Sathwik Tejaswi Madhusudhan","status":"claimed_verified","statusLastChangedAt":"2026-02-17T15:47:50.531Z","hidden":false},{"_id":"68de25d26024653e8a3ed218","user":{"_id":"60ecaa5efee13fee7ada7af4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1637694419284-60ecaa5efee13fee7ada7af4.jpeg","isPro":false,"fullname":"Torsten Scholak","user":"tscholak","type":"user"},"name":"Torsten Scholak","status":"claimed_verified","statusLastChangedAt":"2025-10-17T04:14:39.233Z","hidden":false},{"_id":"68de25d26024653e8a3ed219","name":"Sagar Davasam","hidden":false},{"_id":"68de25d26024653e8a3ed21a","name":"Srinivas Sunkara","hidden":false},{"_id":"68de25d26024653e8a3ed21b","name":"Nicholas Chapados","hidden":false}],"publishedAt":"2025-10-01T17:29:35.000Z","submittedOnDailyAt":"2025-10-06T07:20:15.592Z","title":"Apriel-1.5-15b-Thinker","submittedOnDailyBy":{"_id":"62d913739a5353eef9d7edf3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d913739a5353eef9d7edf3/pRgen2izGJle3ahupOdC7.jpeg","isPro":false,"fullname":"Aman Tiwari","user":"amant555","type":"user"},"summary":"We present Apriel-1.5-15B-Thinker, a 15-billion parameter open-weights\nmultimodal reasoning model that achieves frontier-level performance through\ntraining design rather than sheer scale. Starting from Pixtral-12B, we apply a\nprogressive three-stage methodology: (1) depth upscaling to expand reasoning\ncapacity without pretraining from scratch, (2) staged continual pre-training\nthat first develops foundational text and vision understanding, then enhances\nvisual reasoning through targeted synthetic data generation addressing spatial\nstructure, compositional understanding, and fine-grained perception, and (3)\nhigh-quality text-only supervised fine-tuning on curated instruction-response\npairs with explicit reasoning traces spanning mathematics, coding, science, and\ntool use. Notably, our model achieves competitive results without reinforcement\nlearning or preference optimization, isolating the contribution of our\ndata-centric continual pre-training approach. On the Artificial Analysis\nIntelligence Index, Apriel-1.5-15B-Thinker attains a score of 52, matching\nDeepSeek-R1-0528 despite requiring significantly fewer computational resources.\nAcross ten image benchmarks, its performance is on average within five points\nof Gemini-2.5-Flash and Claude Sonnet-3.7, a key achievement for a model\noperating within single-GPU deployment constraints. Our results demonstrate\nthat thoughtful mid-training 2 design can close substantial capability gaps\nwithout massive scale, making frontier-level multimodal reasoning accessible to\norganizations with limited infrastructure. We release the model checkpoint, all\ntraining recipes, and evaluation protocols under the MIT license to to advance\nopen-source research.","upvotes":121,"discussionId":"68de25d26024653e8a3ed21c","ai_summary":"A 15-billion parameter multimodal reasoning model achieves competitive performance through a progressive training methodology without reinforcement learning, demonstrating efficient use of computational resources.","ai_keywords":["multimodal reasoning model","depth upscaling","staged continual pre-training","synthetic data generation","spatial structure","compositional understanding","fine-grained perception","text-only supervised fine-tuning","reasoning traces","Artificial Analysis Intelligence Index","single-GPU deployment"],"organization":{"_id":"65f4df5de83b55da5d79fbb6","name":"ServiceNow-AI","fullname":"ServiceNow-AI","avatar":"https://cdn-uploads.huggingface.co/production/uploads/63d3095c2727d7888cbb54e2/Uv-Lx8PVGviqokfOyYlCN.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"62d913739a5353eef9d7edf3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62d913739a5353eef9d7edf3/pRgen2izGJle3ahupOdC7.jpeg","isPro":false,"fullname":"Aman Tiwari","user":"amant555","type":"user"},{"_id":"63d3095c2727d7888cbb54e2","avatarUrl":"/avatars/3d437afdf19cb43c6b67a15e4c2955f8.svg","isPro":false,"fullname":"Sathwik Tejaswi Madhusudhan","user":"stm4","type":"user"},{"_id":"64a3bbc378edd17f9652bf3b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64a3bbc378edd17f9652bf3b/XJM8hvtNUxRSMrWvznLHT.jpeg","isPro":false,"fullname":"Amit Kumar Saha","user":"amitsaha","type":"user"},{"_id":"64b935fc49bde5d94836202d","avatarUrl":"/avatars/4ede3a8cf5afc556070edc16c9262cbd.svg","isPro":false,"fullname":"Sourabh Surana","user":"sourabh89","type":"user"},{"_id":"661909cd3b412cdc853f839b","avatarUrl":"/avatars/0658fd156ed521d013dd1b39dda2c30a.svg","isPro":false,"fullname":"sandeep reddy katypally","user":"sandeep-katypally","type":"user"},{"_id":"603c6bf03249b99991dbcbd0","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/603c6bf03249b99991dbcbd0/IZDS9PIg-9b8QleK08Zan.png","isPro":false,"fullname":"Surajit Dasgupta","user":"surajit","type":"user"},{"_id":"642f99079b2484d7d857341b","avatarUrl":"/avatars/01965cc5a5dbe9c08025a51973462a6a.svg","isPro":false,"fullname":"Sai Rajeswar","user":"rajeswarsai","type":"user"},{"_id":"681eca14a888a43c883911b6","avatarUrl":"/avatars/ccbda4533fb9fbee37eb115194cfc019.svg","isPro":false,"fullname":"Sourav Sharma","user":"sourav-snow","type":"user"},{"_id":"67a254f0aad4ef82acf649ff","avatarUrl":"/avatars/dae66fc26030070368867ac7a98c27d2.svg","isPro":false,"fullname":"Tunde Oladipo","user":"TundeAtSN","type":"user"},{"_id":"64895ed451b69a8c82f493a4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64895ed451b69a8c82f493a4/0G3lDWoVWiFuFmPfXPKfQ.jpeg","isPro":false,"fullname":"Seganrasan Subramanian","user":"Seganrasan","type":"user"},{"_id":"6697bc94c15a4f7037afbddb","avatarUrl":"/avatars/7258dfa45781065f8a0dad71ef1059fe.svg","isPro":false,"fullname":"Verma","user":"Abhigy","type":"user"},{"_id":"64e7ac350c47bf287ca307ef","avatarUrl":"/avatars/b36c727cbb983615e2dada2ef769639b.svg","isPro":false,"fullname":"Fanny Riols","user":"Nny042","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":1,"organization":{"_id":"65f4df5de83b55da5d79fbb6","name":"ServiceNow-AI","fullname":"ServiceNow-AI","avatar":"https://cdn-uploads.huggingface.co/production/uploads/63d3095c2727d7888cbb54e2/Uv-Lx8PVGviqokfOyYlCN.png"}}">