United States
IBM
IBM focuses on enterprise AI with Watsonx, offering foundation models, governance tools, and industry-specific AI solutions.
Region
United States
Updated
May 14, 2026
Product coverage
Products from this provider
Model coverage
Models from this provider
Granite 3.3
Granite 3.3 8B (Non-reasoning)
An 8B parameter, instruction-tuned model from IBM's Granite 3.3 series, designed for general-purpose tasks and edge deployment. This non-reasoning variant focuses on efficient instruction following and text generation without extended chain-of-thought processes.
Input / 1M tokens
$0.03
Output tokens/s
382.62
First-token seconds
20.15s
Artificial Analysis Intelligence Index
7
Granite 4.0
Granite 4.0 1B
IBM Granite 4.0 1B is a lightweight, 1-billion parameter model from the Granite series, optimized for fast inference and low-cost deployment. It is designed for edge computing and resource-constrained environments, offering efficient performance for straightforward tasks.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
7.3
Granite 4.0
Granite 4.0 350M
A lightweight, 350-million parameter model from IBM's Granite 4.0 series, designed for efficient inference and deployment on resource-constrained environments.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
6.1
Granite 4.0 H
Granite 4.0 H 1B
A compact, 1-billion parameter model from IBM's Granite 4.0 series, optimized for efficiency and speed. It is designed for deployment in resource-constrained environments or as a lightweight component in larger systems.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
8
Granite 4.0 H
Granite 4.0 H 350M
IBM Granite 4.0 H 350M is a compact AI model from the Granite family, featuring 350 million parameters for efficient inference. It is designed for fast and cost-effective deployment in enterprise scenarios, emphasizing speed and low operational costs.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
5.4
Granite 4.0 H
Granite 4.0 H Small
Granite 4.0 H Small is a compact, efficient model from IBM's Granite family, designed for enterprise applications. It offers a balance of strong reasoning capabilities and fast response times, making it suitable for tasks requiring quick and reliable inference.
Input / 1M tokens
$0.06
Output tokens/s
265.36
First-token seconds
8.77s
Artificial Analysis Intelligence Index
10.8
Granite 4.0
Granite 4.0 Micro
Granite 4.0 Micro is a lightweight, efficient model from IBM's Granite family, optimized for fast inference and deployment in resource-constrained environments. It maintains the enterprise-grade reliability and safety standards of the Granite series while offering a smaller footprint for cost-effective applications.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
7.7
Granite 4.1
Granite 4.1 30B
IBM Granite 4.1 30B is an enterprise-grade, open-source foundation model designed for high performance and reliability. It excels in complex reasoning, code generation, and multilingual understanding, making it suitable for business applications requiring accuracy and security.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
14.7
Granite 4.1
Granite 4.1 3B
IBM provides the Granite family of AI models, including Granite 4.1, which are enterprise-ready, open foundation models designed for coding, reasoning, and multilingual tasks. These models support tool-based instructions and are optimized for robust AI workflows in enterprise settings.
Input / 1M tokens
$0.00
Artificial Analysis Intelligence Index
8.5
Granite 4.1
Granite 4.1 8B
IBM Granite 4.1 8B is an enterprise-grade, compact language model designed for high efficiency and reliability. It is optimized for tasks requiring strong reasoning and coding capabilities while maintaining a low cost and fast inference speed, making it suitable for deployment in resource-constrained or latency-sensitive environments.
Input / 1M tokens
$0.05
Output tokens/s
130.02
First-token seconds
0.43s
Artificial Analysis Intelligence Index
12.4
Discussion

Thinking... Make sure you are connected to GitHub server

