Midm-2.0-Mini-Instruct: benchmark results
Provider: Other.
Unified ELO 1409 ± 32, rank #1146 of 1607 rated models, from 39 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Polar (South Korea) - Economic Axis | 90.03 | Economic-axis ICAT (0-100), the StereoSet idealized context | 91.9 |
| Polar (South Korea) | 90.99 | Total ICAT (0-100), the StereoSet idealized context associat | 83.8 |
| Polar (South Korea) - Sociocultural Axis | 91.96 | Sociocultural-axis ICAT (0-100), the StereoSet idealized con | 75.7 |
| Horangi 4 - KoBBQ | 75 | Accuracy (%) | 27.4 |
| Polar (US) | 73.81 | Total ICAT (0-100), the StereoSet idealized context associat | 27 |
| Polar (US) - Sociocultural Axis | 80.46 | Sociocultural-axis ICAT (0-100), the StereoSet idealized con | 27 |
| Polar (US) - Economic Axis | 67.16 | Economic-axis ICAT (0-100), the StereoSet idealized context | 24.3 |
| Horangi 4 - Ko-HLE | 8 | Accuracy (%) | 18.8 |
| Horangi 4 - Ko-HalluLens (WikiQA) | 10 | Correct answer rate (%) | 13.5 |
| Horangi 4 - HAE-RAE Bench (without Reading Comprehension) | 68.69 | Accuracy (%) | 11.1 |
| Horangi 4 - IFEval-Ko | 70.42 | Instruction-following accuracy (%) | 10.6 |
| Horangi 4 - GLP - Specialized Knowledge | 26 | Score (%) | 10.1 |
Interactive version: theaggregate.ai/model?slug=midm-2-0-mini-instruct · How It Works · Data refreshed daily, snapshot 2026-09-29.