MN-12B-Mag-Mell-R1: benchmark results
inflatebot's multi-stage DARE-TIES merge of six Mistral Nemo 12B fine-tunes, built as a 'best of Nemo' model for creative fiction and roleplay. Provider: Other. Released 2024-09-16. Access: Open.
Unified ELO 1497 ± 1, rank #701 of 1392 rated models, from 11 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard - GPQA | 8.84 | Score | 72.5 |
| Open LLM Leaderboard - BBH | 32.54 | Score | 61.8 |
| Open LLM Leaderboard - MuSR | 11.3 | Score | 58.9 |
| UGI - Writing | 35.82 | Writing Score | 57.9 |
| Open LLM Leaderboard - MATH Level 5 | 12.99 | Score | 57.5 |
| UGI - Willingness (W/10) | 6 | W/10 Score | 54.2 |
| Open LLM Leaderboard - IFEval | 46.13 | Score | 52.1 |
| Open LLM Leaderboard - MMLU-Pro | 27.09 | Score | 49.6 |
| UGI - Natural Intelligence | 21.63 | NatInt Score | 49.3 |
| UGI Leaderboard | 31.14 | UGI Score | 35.2 |
| FMNB Leaderboard | 53 | Score | 18.8 |
Interactive version: theaggregate.ai/model?slug=mn-12b-mag-mell-r1 · How It Works · Data refreshed daily, snapshot 2026-09-05.