G9v3-3B: benchmark results
Provider: AI9Stars. Released 2026-07-23. Access: Open.
Aggregate Rating 1423 ± 43, rank #5162 of 10516 rated models, from 22 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA-Omniscience Hallucination Rate | 11.66 | Hallucination Rate (%) | 99.8 |
| AA Omniscience | -4.37 | Score | 70.1 |
| AA Long Context Reasoning | 41.33 | Accuracy (%) | 34.9 |
| AA Omniscience - Software Engineering (SWE) - HTML | 20 | Accuracy (%) | 22 |
| AA CritPt | 0 | Accuracy (%) | 20.6 |
| AA Omniscience - Software Engineering (SWE) - Swift | 20 | Accuracy (%) | 20.5 |
| AA Humanity's Last Exam | 4.49 | Accuracy (%) | 19.8 |
| AA GPQA Diamond | 43.84 | Accuracy (%) | 19.2 |
| AA Omniscience - Software Engineering (SWE) - C | 17 | Accuracy (%) | 15.5 |
| AA Terminal-Bench 2.1 | 5.99 | Pass@1 (%) | 14 |
| AA Omniscience - Software Engineering (SWE) - Julia | 0 | Accuracy (%) | 10.8 |
| AA Omniscience - Software Engineering (SWE) - Rust | 24 | Accuracy (%) | 9 |
Interactive version: theaggregate.ai/model?slug=g9v3-3b · How It Works · Data refreshed daily, snapshot 2026-11-21.