Gemma2-Swallow-27B: benchmark results
Provider: Other.
Aggregate Rating 1643 ± 47, rank #672 of 10916 rated models, from 2 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Japanese G2P (Parse Mode) | 5.74 | Kana character error rate (%; can exceed 100) in parse mode | 66.7 |
| Benchmarking Large Language Models for Graphem | 17.53 | Direct (self-reported) | 55.6 |
Interactive version: theaggregate.ai/model?slug=gemma2-swallow-27b · How It Works · Data refreshed daily, snapshot 2026-11-08.