Apriel-1.5-15b-Thinker
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911187983138816 |
|---|---|
| author | Radhakrishna, Shruthan Tiwari, Aman Shukla, Aanjaneya Hashemi, Masoud Maheshwary, Rishabh Malay, Shiva Krishna Reddy Mehta, Jash Pattnaik, Pulkit Mittal, Saloni Slimi, Khalil Ogueji, Kelechi Oladipo, Akintunde Parikh, Soham Bamgbose, Oluwanifemi Liang, Toby Masry, Ahmed Mahajan, Khyati Mudumba, Sai Rajeswar Yadav, Vikas Madhusudhan, Sathwik Tejaswi Scholak, Torsten Davasam, Sagar Sunkara, Srinivas Chapados, Nicholas |
| author_facet | Radhakrishna, Shruthan Tiwari, Aman Shukla, Aanjaneya Hashemi, Masoud Maheshwary, Rishabh Malay, Shiva Krishna Reddy Mehta, Jash Pattnaik, Pulkit Mittal, Saloni Slimi, Khalil Ogueji, Kelechi Oladipo, Akintunde Parikh, Soham Bamgbose, Oluwanifemi Liang, Toby Masry, Ahmed Mahajan, Khyati Mudumba, Sai Rajeswar Yadav, Vikas Madhusudhan, Sathwik Tejaswi Scholak, Torsten Davasam, Sagar Sunkara, Srinivas Chapados, Nicholas |
| contents | We present Apriel-1.5-15B-Thinker, a 15-billion parameter open-weights multimodal reasoning model that achieves frontier-level performance through training design rather than sheer scale. Starting from Pixtral-12B, we apply a progressive three-stage methodology: (1) depth upscaling to expand reasoning capacity without pretraining from scratch, (2) staged continual pre-training that first develops foundational text and vision understanding, then enhances visual reasoning through targeted synthetic data generation addressing spatial structure, compositional understanding, and fine-grained perception, and (3) high-quality text-only supervised fine-tuning on curated instruction-response pairs with explicit reasoning traces spanning mathematics, coding, science, and tool use. Notably, our model achieves competitive results without reinforcement learning or preference optimization, isolating the contribution of our data-centric continual pre-training approach. On the Artificial Analysis Intelligence Index, Apriel-1.5-15B-Thinker attains a score of 52, matching DeepSeek-R1-0528 despite requiring significantly fewer computational resources. Across ten image benchmarks, its performance is on average within five points of Gemini-2.5-Flash and Claude Sonnet-3.7, a key achievement for a model operating within single-GPU deployment constraints. Our results demonstrate that thoughtful mid-training 2 design can close substantial capability gaps without massive scale, making frontier-level multimodal reasoning accessible to organizations with limited infrastructure. We release the model checkpoint, all training recipes, and evaluation protocols under the MIT license to to advance open-source research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_01141 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Apriel-1.5-15b-Thinker Radhakrishna, Shruthan Tiwari, Aman Shukla, Aanjaneya Hashemi, Masoud Maheshwary, Rishabh Malay, Shiva Krishna Reddy Mehta, Jash Pattnaik, Pulkit Mittal, Saloni Slimi, Khalil Ogueji, Kelechi Oladipo, Akintunde Parikh, Soham Bamgbose, Oluwanifemi Liang, Toby Masry, Ahmed Mahajan, Khyati Mudumba, Sai Rajeswar Yadav, Vikas Madhusudhan, Sathwik Tejaswi Scholak, Torsten Davasam, Sagar Sunkara, Srinivas Chapados, Nicholas Artificial Intelligence We present Apriel-1.5-15B-Thinker, a 15-billion parameter open-weights multimodal reasoning model that achieves frontier-level performance through training design rather than sheer scale. Starting from Pixtral-12B, we apply a progressive three-stage methodology: (1) depth upscaling to expand reasoning capacity without pretraining from scratch, (2) staged continual pre-training that first develops foundational text and vision understanding, then enhances visual reasoning through targeted synthetic data generation addressing spatial structure, compositional understanding, and fine-grained perception, and (3) high-quality text-only supervised fine-tuning on curated instruction-response pairs with explicit reasoning traces spanning mathematics, coding, science, and tool use. Notably, our model achieves competitive results without reinforcement learning or preference optimization, isolating the contribution of our data-centric continual pre-training approach. On the Artificial Analysis Intelligence Index, Apriel-1.5-15B-Thinker attains a score of 52, matching DeepSeek-R1-0528 despite requiring significantly fewer computational resources. Across ten image benchmarks, its performance is on average within five points of Gemini-2.5-Flash and Claude Sonnet-3.7, a key achievement for a model operating within single-GPU deployment constraints. Our results demonstrate that thoughtful mid-training 2 design can close substantial capability gaps without massive scale, making frontier-level multimodal reasoning accessible to organizations with limited infrastructure. We release the model checkpoint, all training recipes, and evaluation protocols under the MIT license to to advance open-source research. |
| title | Apriel-1.5-15b-Thinker |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2510.01141 |