TPSTokens Per Second

Granite 4.2

Ibm🇺🇸 United States
active
Compare with other models →
Context window131K tokens

Version History

4.2minor

IBM released Granite 4.2, its first reasoning-focused LLM family in 3B, 8B, and 30B dense sizes, pre-trained on ~15T tokens with a five-phase pipeline and post-trained with multi-stage RL including agentic tool-use training for the 8B and 30B models.

Coverage

researchIbm

IBM Releases Granite 4.2, Its First Reasoning-Focused LLM Family in 3B, 8B, and 30B Sizes

IBM has published a technical breakdown of Granite 4.2, its first dense, decoder-only reasoning model family, released in 3B, 8B, and 30B sizes. The models are pre-trained on roughly 15 trillion tokens, extended to a 512K-token context window, and post-trained with a multi-stage RL pipeline that includes agentic tool-use training for the 8B and 30B variants.

3 min read
IBMGraniteopen-source-llm