IBM Granite 4.0 Tiny Preview
Overview
A preliminary version of the smallest model in the upcoming Granite 4.0 family, released May 2025. It utilizes a novel hybrid Mamba-2/Transformer, fine-grained mixture of experts (MoE) architecture (7B total parameters, 1B active at inference). This preview version is partially trained (2.5T tokens) but demonstrates significant memory efficiency and performance potential, validated for at least 128K context length without positional encoding.
IBM Granite 4.0 Tiny Preview was released on May 2, 2025.
Performance
Timeline
Specifications
Benchmarks
IBM Granite 4.0 Tiny Preview Performance Across Datasets
Scores sourced from the model's scorecard, paper, or official blog posts
Pricing
Pricing, performance, and capabilities for IBM Granite 4.0 Tiny Preview across different providers:
API Access
API Access Coming Soon
API access for IBM Granite 4.0 Tiny Preview will be available soon through our gateway.
Recent Posts
Recent Reviews
FAQ
Common questions about IBM Granite 4.0 Tiny Preview
