Redis's langcache-embed-v3-small, JetBrains' Mellum2-12B-A2.5B-Thinking, and LightOn's LightOnOCR-2-1B are now deployable on AWS SageMaker JumpStart across all regions where the service operates. You can now spin up semantic caching for LLMs, code-focused reasoning with 131K context length, and multilingual document OCR—all with a few clicks, no brittle pipelines required.
source: [aws/whats-new]
Z.ai's GLM-5.2 FP8, NVIDIA's Nemotron-Nano-12B-v2, and Z.ai's GLM-OCR are now available on Amazon SageMaker JumpStart, letting you deploy specialized models for long-context coding tasks, efficient reasoning, and document understanding with just a few clicks. GLM-5.2 FP8 handles full software development workflows with its 1M-token context window, Nemotron-Nano-12B-v2 delivers 6x faster inference while maintaining accuracy, and GLM-OCR extracts structured data from complex PDFs and handwritten documents—all ready to scale on AWS infrastructure.
source: [aws/whats-new]