Granite 4.2 8B

Granite 4.2 8B is a chat-oriented language model from IBM, part of the Granite 4.2 generation of the company's Granite series. At roughly 8 billion parameters, it sits in the small-model tier of that family — the size class typically chosen when a deployment needs conversational quality at a lower compute footprint than a mid-size or large model, rather than maximum capability on hard reasoning tasks. In measured serving performance, third-party benchmarking from Artificial Analysis records Granite 4.2 8B at approximately 150 output tokens per second with a time to first token of about 231 ms. That combination — sub-quarter-second first-token latency and triple-digit generation throughput — is the main quantitative signal we hold on this model, and it is the characteristic most relevant to interactive chat, streaming UIs, and high-request-volume pipelines. Other specifications for this entry, including context window length, input modalities, and tool-calling support, are not tracked in our database; check IBM's model documentation or your serving provider's API reference for those details. In practice, models in this size and family are commonly used for enterprise text tasks — summarization, drafting, classification, retrieval-augmented question answering, and internal assistants — where predictable latency and lower per-request cost matter more than frontier benchmark scores. Buyers evaluating Granite 4.2 8B typically compare it against other 7B–9B chat models, and against larger models in IBM's own Granite lineup when a task turns out to need more capability than an 8B model provides.