qwen2.5-bakeneko-32B-instruct-v2: benchmark results
Rinna's Qwen2.5 Bakeneko 32B Instruct v2, enhanced via Chat Vector from their QwQ Bakeneko 32B reasoning model, giving reasoning-like Japanese performance without an explicit reasoning process. Provider: rinna. Released 2025-03-16.
Aggregate Rating 1565 ± 22, rank #562 of 1656 rated models, from 146 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Nejumi 4 - BFCL - Relevance Detection | 100 | Accuracy (%) | 98.6 |
| Open Japanese LLM - Mbpp Pylint Check | 99 | Score (%) | 98.2 |
| Open Japanese LLM - Jnli Exact Match | 89.77 | Score (%) | 96.3 |
| Open Japanese LLM - CG | 63.86 | Score (%) | 94.8 |
| Nejumi 4 - BFCL - Non-Live AST | 81.85 | Accuracy (%) | 94.4 |
| Open Japanese LLM - NLI | 81.3 | Score (%) | 93.4 |
| Open Japanese LLM - Jamp Exact Match | 70.69 | Score (%) | 92.9 |
| Open Japanese LLM Leaderboard | 66.35 | Average Score (%) | 92.3 |
| Open Japanese LLM - Wiki PAS SET F1 | 10.65 | Score (%) | 91.3 |
| Open Japanese LLM - Xlsum JA Bert Score JA F1 | 70.76 | Score (%) | 90.7 |
| Open Japanese LLM - MT | 85.16 | Score (%) | 90 |
| Open Japanese LLM - Wiki Dependency SET F1 | 42.16 | Score (%) | 89.9 |
Interactive version: theaggregate.ai/model?slug=qwen2-5-bakeneko-32b-instruct-v2 · How It Works · Data refreshed daily, snapshot 2026-10-20.