Granite 4.2 8B: benchmark results
IBM's Granite 4.2 8B, the mid-size reasoning model in the Granite 4.2 family, with native chain-of-thought and switchable full, non-thinking and low-effort thinking modes. Provider: IBM. Released 2026-08-25. Access: Open.
Aggregate Rating 1543 ± 28, rank #635 of 1656 rated models, from 113 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA-Omniscience Hallucination Rate | 31.97 | Hallucination Rate (%) | 91 |
| AA-Omniscience Index - Law | -8.1 | Omniscience Index | 72.3 |
| AA-Omniscience Index - Business | -9.1 | Omniscience Index | 70.1 |
| AA-Omniscience Index - Humanities & Social Sciences | -7.8 | Omniscience Index | 69.3 |
| AA Omniscience | -17.17 | Score | 59.8 |
| AA-Omniscience Index - Science, Engineering & Mathematics | -13.4 | Omniscience Index | 57.9 |
| AA-Omniscience Index - Software Engineering (SWE) - HTML | -10 | Omniscience Index | 55 |
| AA-Omniscience Index - Health | -29.5 | Omniscience Index | 54.7 |
| AA-Omniscience Index - Software Engineering (SWE) - Dart | -32 | Omniscience Index | 52.8 |
| AA-Omniscience Index - Software Engineering (SWE) - Julia | -44 | Omniscience Index | 50.2 |
| LLM Stats (HMMT25; February 2025 (Granite evaluation)) | 78.33 | Score (%) | 50 |
| AA-Omniscience Index - Software Engineering (SWE) - JavaScript | -22.73 | Omniscience Index | 49.6 |
Interactive version: theaggregate.ai/model?slug=granite-4-2-8b · How It Works · Data refreshed daily, snapshot 2026-10-20.