Polka-Mistral-7B-SFT: benchmark results
Provider: Other.
Aggregate Rating 1332 ± 36, rank #8130 of 10916 rated models, from 9 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MT-Bench PL - Math | 2.95 | Judge Score (0-10) | 20.4 |
| MT-Bench PL - STEM | 6.8 | Judge Score (0-10) | 20.4 |
| MT-Bench PL - Writing | 5.25 | Judge Score (0-10) | 16.3 |
| MT-Bench PL - Extraction | 5.25 | Judge Score (0-10) | 14.3 |
| MT-Bench PL - Overall | 4.43 | Judge Score (0-10) | 14.3 |
| MT-Bench PL - Roleplay | 4.9 | Judge Score (0-10) | 14.3 |
| MT-Bench PL - Humanities | 5.6 | Judge Score (0-10) | 12.2 |
| MT-Bench PL - Coding | 2.95 | Judge Score (0-10) | 8.2 |
| MT-Bench PL - Reasoning | 2.45 | Judge Score (0-10) | 8.2 |
Interactive version: theaggregate.ai/model?slug=polka-mistral-7b-sft · How It Works · Data refreshed daily, snapshot 2026-11-08.