NuminaMath-7B-TIR: benchmark results
Provider: Numina & Hugging Face. Released 2024-07-04. Access: Open.
Aggregate Rating 1337 ± 43, rank #8042 of 10916 rated models, from 15 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Korean LLM - Ko-GPQA-Diamond | 26.77 | Normalized accuracy (%) | 88.9 |
| Open LLM Leaderboard - MATH Level 5 | 16.09 | Score | 63.9 |
| Open Korean LLM - Ko-IFEval | 29.26 | Strict accuracy, prompt/instruction mean (%) | 50.9 |
| Open Korean LLM - KorNAT-Helpful | 46.68 | Normalized accuracy (%) | 48.3 |
| Open Korean LLM - KorNAT-CKA | 23.2 | Normalized accuracy (%) | 43.5 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 9.48 | Exact match, flexible extract (%) | 32.7 |
| Open Korean LLM - KorNAT-Harmless | 61.57 | Normalized accuracy (%) | 31.1 |
| Open LLM Leaderboard - MMLU-Pro | 27.33 | Score | 29.4 |
| Open Korean LLM Leaderboard | 31.55 | Average Score (%) | 27.9 |
| Open LLM Leaderboard - BBH | 41.44 | Score | 26.4 |
| Open Korean LLM - Ko-EQ-Bench | 0 | EQ-Bench score | 24.5 |
| Open LLM Leaderboard - IFEval | 27.56 | Score | 24.5 |
Interactive version: theaggregate.ai/model?slug=numinamath-7b-tir · How It Works · Data refreshed daily, snapshot 2026-11-08.