PrismML
Follow
Visit website
Models
-
Ternary Bonsai Image 4B is PrismML’s open-weight compressed image-generation model that trades slightly larger size for stronger visual quality and prompt fidelity.NewImageReleased 25d ago
-
1-bit Bonsai Image 4B is PrismML’s open-weight compressed image-generation model built for local diffusion inference on laptops, phones, and other memory-limited devices.NewImageReleased 25d ago
-
1-bit Bonsai 1.7B is PrismML’s smallest Bonsai model, built for extremely fast on-device inference. With a memory footprint of just 0.24GB, PrismML says it can reach 130 tokens per second on an iPhone 17 Pro Max, making it a lightweight edge model aimed at mobile and highly power-constrained applications.NewTextReleased 2mo ago
-
1-bit Bonsai 4B is PrismML’s mid-size ultra-efficient model designed for fast local inference. It uses just 0.57GB of memory and is positioned as a strong balance of speed, accuracy, and energy efficiency, with PrismML reporting throughput of 132 tokens per second on an M4 Pro for workloads that need both responsiveness and capable language performance.NewTextReleased 2mo ago
-
1-bit Bonsai 8B is PrismML’s flagship ultra-efficient language model, built with 1-bit weights for edge and real-time use. It needs only 1.15GB of memory and is aimed at robotics, real-time agents, and on-device computing, with PrismML claiming a 14x smaller footprint, 8x faster runtime, and 5x better energy efficiency than standard full-precision 8B models.NewTextReleased 2mo ago
MongoDB - Build AI That Scales
