granite-3.1-3B-a800m-instruct: benchmark results
Provider: IBM. Released 2024-12-18. Access: Open.
Aggregate Rating 1313 ± 18, rank #1452 of 1656 rated models, from 100 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Enkrypt AI - Toxicity Risk | 0.23 | Risk Score | 93.7 |
| Enkrypt AI - Jailbreak Risk | 4.41 | Risk Score | 80.1 |
| Open LLM Leaderboard - IFEval | 55.16 | Score | 66.7 |
| Open Japanese LLM - Xlsum JA Rouge1 | 27.56 | Score (%) | 65.4 |
| Open Japanese LLM - Xlsum JA RougeLsum | 23.19 | Score (%) | 63.7 |
| Enkrypt AI - Safety Risk | 24.95 | Risk Score | 57.7 |
| Open LLM Leaderboard - MATH Level 5 | 11.4 | Score | 51.8 |
| Open Japanese LLM - Xlsum JA Bert Score JA F1 | 68.73 | Score (%) | 45.4 |
| Open PL LLM - EQ-Bench | 23.94 | EQ-Bench Score | 44.9 |
| Open LLM Leaderboard - GPQA | 28.86 | Score | 43.2 |
| TelcoAgent-Bench - Resolution Accuracy (English) | 85.5 | Resolution accuracy: mean cosine similarity x100 (-100 to 10 | 42.9 |
| Open Japanese LLM - SUM | 9.2 | Score (%) | 42.8 |
Interactive version: theaggregate.ai/model?slug=granite-3-1-3b-a800m-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.