Granite 4.0 1B: benchmark results
IBM's Granite 4.0 1B instruct model, fine-tuned from Granite-4.0-1B-Base with supervised fine-tuning, reinforcement learning and model merging, supporting 12 languages. Provider: IBM. Released 2025-10-28. Access: Open.
Aggregate Rating 1367 ± 36, rank #1343 of 1656 rated models, from 60 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open PL LLM - PolEmo2-IN (multiple choice, 5-shot) | 72.58 | Accuracy (%) | 35.6 |
| Open PL LLM - PolEmo2-IN (generative, 5-shot) | 68.7 | Accuracy (%) | 29.8 |
| AA Omniscience - Software Engineering (SWE) - Go | 10 | Accuracy (%) | 28.2 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 10.2 | Accuracy (%) | 27.8 |
| Tau2-Bench Telecom | 22.81 | Success Rate (%) | 25.7 |
| AA Humanity's Last Exam | 4.82 | Accuracy (%) | 25 |
| AA TAU-2 Bench | 22.81 | Accuracy (%) | 24.9 |
| AA CritPt | 0 | Accuracy (%) | 22 |
| AA Omniscience - Software Engineering (SWE) - Java | 9.09 | Accuracy (%) | 20.8 |
| CritPt | 0 | Accuracy (imported) | 14.2 |
| AA Omniscience - Software Engineering (SWE) - R | 2 | Accuracy (%) | 14 |
| AA Long Context Reasoning | 6 | Accuracy (%) | 12.1 |
Interactive version: theaggregate.ai/model?slug=granite-4-0-1b · How It Works · Data refreshed daily, snapshot 2026-10-20.