granite-3.1-8B-instruct: benchmark results
IBM's 8B Apache-2.0 enterprise model with a 128K context, from the Granite 3.1 release of December 2024. Provider: IBM. Released 2024-12-18. Access: Open.
Aggregate Rating 1436 ± 18, rank #1069 of 1656 rated models, from 135 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EVALITA - faq | 54.7 | CPS | 95.7 |
| EVALITA - summarization-fanpage | 32.88 | CPS | 93.6 |
| Enkrypt AI - Toxicity Risk | 0.32 | Risk Score | 93.2 |
| Open Portuguese LLM - FaQuAD NLI | 81.51 | Macro F1 (%) | 92.2 |
| Open LLM Leaderboard - MuSR | 47.07 | Score | 91.7 |
| Open Japanese LLM - Xlsum JA Bleu JA | 3.15 | Score (%) | 86.1 |
| Open LLM Leaderboard - IFEval | 72.08 | Score | 85.9 |
| Open Portuguese LLM - ASSIN2 RTE | 92.63 | Macro F1 (%) | 81.2 |
| Open Japanese LLM - Xlsum JA Rouge1 | 29.87 | Score (%) | 80.2 |
| EVALITA - relation-extraction | 39.91 | CPS | 78.7 |
| Enkrypt AI - Jailbreak Risk | 4.67 | Risk Score | 78.1 |
| Open Japanese LLM - Xlsum JA RougeLsum | 24.85 | Score (%) | 75.7 |
Interactive version: theaggregate.ai/model?slug=granite-3-1-8b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.