The Year in AI Papers
The year/Independent research

Paper 2510.01141

Apriel-1.5-15b-Thinker

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, et al.

Published
Oct 2025
Research lab
Independent
Citations
2
GitHub
Not linked

01 In brief

Summary

Apriel-1.5-15B-Thinker is a 15-billion-parameter open-weights multimodal reasoning model developed by ServiceNow's SLAM Lab.

It achieves frontier-level performance through a data-centric mid-training approach rather than scale, starting from Pixtral-12B.

The methodology involves three stages: depth upscaling to expand reasoning capacity, staged continual pretraining (CPT) that first builds foundational text and vision understanding then enhances visual reasoning with synthetic data targeting spatial structure, compositionality, and fine-grained perception, and high-quality supervised fine-tuning (SFT) on curated instruction-response pairs with explicit reasoning traces.

The model attains a score of 52 on the Artificial Analysis Intelligence Index, matching DeepSeek-R1-0528 with fewer computational resources.

Across ten image benchmarks, it averages within five points of Gemini-2.5-Flash and Claude Sonnet-3.7, operating within single-GPU constraints.

Key results include 87.5% on AIME2025, 62% on IFBench, 68% on τ2-Bench Telecom, and 70.22% on MMMU.

The model is released under the MIT license with training recipes and evaluation protocols, demonstrating that thoughtful mid-training can close capability gaps without massive scale, making frontier-level multimodal reasoning accessible to organizations with limited infrastructure.

02 From the paper

Abstract

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.