Alibaba’s Qwen leads with 29 additions, followed by NVIDIA at 17, Mistral at 7, Google at 6, and IBM at 4.

Amazon SageMaker JumpStart onboarded more than 100 new models from 30 distinct providers between January and September 2026. The catalog now spans reasoning, mixture-of-experts, agentic, code, speech, embedding, and reranking models. Teams reach the library through the SageMaker Studio console, the SageMaker Python SDK, or a Hugging Face deep link integration.

JumpStart is a curated hub inside Amazon SageMaker AI that holds model cards, solution templates, and example notebooks. Each card lists architecture details, benchmarks, and licensing terms. Alibaba’s Qwen contributed 29 models, NVIDIA added 17 including the Cosmos3 world model family, Mistral added 7, Google added 6 Gemma 4 variants, and IBM added 4 Granite 4.0 entries.

Teams move from model selection to a real-time inference endpoint with pre-configured settings tuned for low cost, low latency, throughput, or balance. No manual container or instance setup is required. Customization runs from the console or SDK using low-rank adaptation, supervised fine-tuning, direct preference optimization, and reinforcement learning with verifiable rewards, and serverless options cut infrastructure management.

Catalog growth tracks the release cadence of open source labs, with the Qwen, Nemotron, and Gemma families accounting for most 2026 additions. Operators should track which models gain fine-tuning and serverless support, because customization depth matters more than raw model count. Model cards with benchmarks and licensing detail give procurement teams a faster path to approval.

What matters

• The catalog added more than 100 models across 30 providers between January and September 2026.
• Builders can deploy any listed model to a SageMaker endpoint without configuring containers or instances.
• Watch whether fine-tuning options like LoRA, DPO, and RLVR expand alongside the base catalog.

Why it matters

Watch whether fine-tuning options like LoRA, DPO, and RLVR expand alongside the base catalog.

This GenAI News article was prepared in original wording using reporting and materials published by AWS Machine Learning Blog. Source reference: https://aws.amazon.com/blogs/machine-learning/amazon-sagemaker-jumpstart-2026-year-to-date-in-review/.

Drafted by the GenAI News review pipeline.