Apriel-1.6-15b-Thinker: Cost-efficient Frontier Multimodal Performance

ServiceNow releases Apriel-1.6-15b-Thinker, a 15-billion parameter multimodal reasoning model that achieves state-of-the-art performance comparable to models 10x its size. The model scores 57 on the Artificial Analysis Index, outperforming Gemini 2.5 Flash and Claude Haiku 4.5 while reducing reasoning token usage by over 30% compared to its predecessor. Training involved depth-upscaling, continual pretraining on synthetic data and image-text pairs, supervised fine-tuning on 2.4M high-quality samples, and multi-stage reinforcement learning using Group Sequence Policy Optimization. The model excels across text and vision benchmarks including function calling, instruction following, math, coding, and visual reasoning tasks, positioning it in the cost-efficient frontier for enterprise deployment.