Granite 4.2 8B: benchmark results

Provider: IBM. Released 2026-08-25. Access: Open.

Unified ELO 1544 ± 1, rank #464 of 1392 rated models, from 36 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience-17.17Score66.7
Artificial Analysis Intelligence Index13.94Intelligence Index56.1
AA CritPt0.29Accuracy (%)53.2
AA Humanity's Last Exam9.68Accuracy (%)53.1
LLM Stats (HMMT25)78.33Score (%)51.9
BenchmarkList ECI111.95Capability Index (ECI)50.6
LM Market Cap LMC Score54LMC Score (0-100)48.9
LLM Stats Score19.76LLM Stats Score (conservative rating)46.4
AA Long Context Reasoning45Accuracy (%)43.9
Chatbot Arena (Text - Coding)1364Arena Score40.1
AA GPQA Diamond63.13Accuracy (%)38.3
ZeroEval GPQA Diamond64.14GPQA Diamond Score32.9

Interactive version: theaggregate.ai/model?slug=granite-4-2-8b · How It Works · Data refreshed daily, snapshot 2026-09-05.