Rombos-LLM-V2.5-Qwen-32B: benchmark results
A Qwen-family base/instruct TIES merge released by rombodawg using the creator's continuous finetuning recipe. It remains distinct from the upstream Qwen releases. Provider: rombodawg. Access: Open.
Unified ELO 1629 ± 27, rank #353 of 1636 rated models, from 65 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - MMLU-Pro | 59.16 | Score | 99.8 |
| Open LLM Leaderboard - BBH | 70.46 | Score | 99.2 |
| Open LLM Leaderboard - GPQA | 39.68 | Score | 98.8 |
| Open LLM Leaderboard - MuSR | 50.34 | Score | 98.7 |
| Open Portuguese LLM - ASSIN2 RTE | 94.69 | Macro F1 (%) | 98 |
| Open Portuguese LLM - BLUEX | 78.58 | Accuracy (%) | 97.8 |
| Open Portuguese LLM - ENEM | 84.81 | Accuracy (%) | 97.8 |
| Open Japanese LLM - MR | 94.4 | Score (%) | 96.8 |
| Open LLM Leaderboard - MATH Level 5 | 49.55 | Score | 96.5 |
| Open Japanese LLM - Jnli Exact Match | 89.77 | Score (%) | 96.3 |
| Open Portuguese LLM - OAB Exams | 63.05 | Accuracy (%) | 95.6 |
| Open Japanese LLM - Jsick Exact Match | 85.79 | Score (%) | 95.4 |
Interactive version: theaggregate.ai/model?slug=rombos-llm-v2-5-qwen-32b · How It Works · Data refreshed daily, snapshot 2026-10-11.