paper-with-me

Papers

Apriel-1.5-15b-Thinker

2025-10-01 · Shruthan Radhakrishna, Aman Tiwari, Aanjaneya Shukla, Masoud Hashemi, Rishabh Maheshwary, Shiva Krishna Reddy Malay, Jash Mehta, Pulkit Pattnaik, Saloni Mittal, Khalil Slimi, Kelechi Ogueji, Akintunde Oladipo, Soham Parikh, Oluwanifemi Bamgbose, Toby Liang, Ahmed Masry, Khyati Mahajan, Sai Rajeswar Mudumba, Vikas Yadav, Sathwik Tejaswi Madhusudhan, Torsten Scholak, Sagar Davasam, Srinivas Sunkara, Nicholas Chapados arxiv

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.

📄 PDF Abstract BibTeX arXiv:2510.01141

Code (0)

등록된 구현이 없습니다.

Tasks

Synthetic Data GenerationReinforcement LearningMultimodal ReasoningVisual Reasoning

Similar Papers 제목 키워드 기반

Apriel-Nemotron-15B-Thinker

2025-08-13 · Shruthan Radhakrishna, Soham Parikh, Gopal Sarda, Anil Turkkan 외 arxiv

While large language models (LLMs) have achieved remarkable reasoning capabilities across domains like code, math and other enterprise tasks, their significant memory and computational costs often preclude their use in p…

Reinforcement Learning

Apriel-H1: Towards Efficient Enterprise Reasoning Models

2025-11-04 · Oleksiy Ostapenko, Luke Kumar, Raymond Li, Denis Kocetkov 외 arxiv

Large Language Models (LLMs) achieve remarkable reasoning capabilities through transformer architectures with attention mechanisms. However, transformers suffer from quadratic time and memory complexity in the attention …

AprielGuard

2025-12-23 · Jaykumar Kasundra, Anjaneya Praharaj, Sourabh Surana, Lakshmi Sirisha Chodisetty 외 arxiv

Safeguarding large language models (LLMs) against unsafe or adversarial behavior is critical as they are increasingly deployed in conversational and agentic settings. Existing moderation tools often treat safety risks (e…

Super Apriel: One Checkpoint, Many Speeds

2026-04-21 · SLAM Labs, :, Oleksiy Ostapenko, Raymond Li 외 arxiv

We release Super Apriel, a 15B-parameter supernet in which every decoder layer provides four trained mixer choices -- Full Attention (FA), Sliding Window Attention (SWA), Kimi Delta Attention (KDA), and Gated DeltaNet (G…

Apriel-1.5-OpenReasoner: RL Post-Training for General-Purpose and Efficient Reasoning

2026-04-02 · Rafael Pardinas, Ehsan Kamalloo, David Vazquez, Alexandre Drouin arxiv

Building general-purpose reasoning models using reinforcement learning with verifiable rewards (RLVR) across diverse domains has been widely adopted by frontier open-weight models. However, their training recipes and dom…

Reinforcement LearningInstruction FollowingCode Generation