Granite 3.3 8B (Non-reasoning) — benchmark results
Provider: IBM. Released 2025-02-26. Access: Open.
Unified ELO 1305 ± 31, rank #1520 of 1776 rated models, from 37 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA Omniscience - Software Engineering (SWE) - R | 6 | Accuracy (%) | 30.9 |
| CritPt | 0 | Accuracy (self-reported) | 30.3 |
| AA CritPt | 0 | Accuracy (%) | 27.1 |
| AA MATH-500 | 66.47 | Accuracy (%) | 24.8 |
| AA Omniscience - Software Engineering (SWE) - Go | 10 | Accuracy (%) | 24.7 |
| AA Omniscience - Software Engineering (SWE) - TypeScript | 11.11 | Accuracy (%) | 22.8 |
| AA Omniscience - Software Engineering (SWE) - PHP | 14 | Accuracy (%) | 22.5 |
| AA-LCR | 4.3 | Score (self-reported) | 20.9 |
| AA Omniscience - Software Engineering (SWE) - Python | 11 | Accuracy (%) | 19.9 |
| AA Omniscience - Software Engineering (SWE) - C | 21 | Accuracy (%) | 18.4 |
| AA Omniscience - Software Engineering (SWE) - HTML | 18 | Accuracy (%) | 17.5 |
| AA Humanity's Last Exam | 4.22 | Accuracy (%) | 16.2 |
Interactive version: theaggregate.ai/model?slug=granite-3-3-8b-non-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-22.