Sold by: IBM Data and AI
Deployed on AWS
IBM Granite 4.2 30B is a dense, 30-billion-parameter reasoning model with a 128K-token context window, built for reasoning, code generation, tool calling and agentic workflows. Released under the Apache 2.0 license.
Overview
IBM Granite 4.2 30B is the flagship reasoning model in the Granite 4.2 family: a dense, decoder-only transformer with about 30 billion parameters (bfloat16), using grouped-query attention, RoPE, SwiGLU and RMSNorm. It was post-trained with supervised fine-tuning followed by reinforcement learning across math, code, science, instruction following, tool use, chat and structured output.
The model reasons before it answers. It supports full thinking (default), non-thinking, and low-effort modes, selected through the chat template. When it calls tools, it reasons about which function to invoke before making the call.
This SageMaker model runs on vLLM and exposes an OpenAI-style chat API. It serves the full 128K-token context in bfloat16 on a single Amazon EC2 instance.
Suggested uses: reasoning tasks, code generation and coding agents, tool calling and multi-step agentic workflows, retrieval-augmented generation, summarization, and multilingual dialog.
The model is released under Apache 2.0, which permits commercial and research use.
Highlights
- Reasoning and agentic coding. Built-in chain-of-thought with three thinking modes. IBM reports scores such as SWE-Bench Verified 57, AIME25 89.17 and BFCL v4 61.39 for the 30B model.
- Tool calling and long context. Native function calling in the OpenAI function schema, with reasoning returned in a separate field, and a 128K-token context window.
- Multilingual. Tested on 12 languages: English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch and Chinese. Users can fine-tune for additional languages.
Details
Sold by
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
This product is available free of charge. Free subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Vendor refund policy
This model is provided by IBM completely free of charge. No payment is required to use it. Therefore, there are no purchases to refund.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
The IBM Granite 4.2 30B model is released under the Apache 2.0 license. It is a dense, decoder-only transformer with about 30 billion parameters, post-trained with supervised fine-tuning and reinforcement learning for reasoning, code generation, tool calling and agentic workflows. It reasons before it answers, with selectable full-thinking, non-thinking and low-effort modes, and it supports function calling in the OpenAI tool schema. The model handles a 128K-token context window and was tested in 12 languages: English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch and Chinese. This SageMaker model runs on vLLM with an OpenAI-style chat API for real-time inference.
Additional details
Inputs
- Summary
The model can be invoked by passing a prompt. Please see the sample notebook for details.
- Input MIME type
- application/json
https://github.com/ibm-granite-community/SageMaker/blob/main/granite-4-2-30b/real_time_sample_input_data.json
https://github.com/ibm-granite-community/SageMaker/blob/main/granite-4-2-30b/batch_sample_input_data.json
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
messages | list of chat messages (role, content) | application/json | Yes |
temperature | Randomness of sampling. IBM recommends 1.0. | number | No |
max_tokens | Maximum tokens to generate. Thinking tokens count toward this limit. | integer | No |
top_p | Nucleus sampling threshold. IBM recommends 0.95. | number | No |
tools | Functions the model may call, in the OpenAI function schema. | array | No |
tool_choice | auto, none or required. | string | No |
chat_template_kwargs | {"enable_thinking": false} turns thinking off or {"enable_thinking": true, "low_effort": true} prompts the model in low effort thinking. | object | No |
Support
Vendor support
Support is not provided for this product.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
IBM® Sterling Order Management helps you to deliver a seamless order management experience across all channels. With Sterling Order Management, you can enhance cross-channel efficiencies, and improve and streamline complex product order management and fulfillment operations across your commerce systems.
IBM's Granite 3.0 8B Instruct is an 8B-parameter AI model for enterprise use, excelling in multilingual and code tasks; Apache 2.0 licensed.
IBM Granite 4.0 H micro is a hybrid dense model with 3B total parameters, offering long-context efficiency, compact inference, and open licensing.
IBM's Granite 3.0 2B Instruct is a 2B-parameter AI model for enterprise use, excelling in multilingual and code tasks; Apache 2.0 licensed.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.