polyglot-math-4x7B: benchmark results
A multilingual mixture-of-experts merge released by macadeliccc combining meta-math/MetaMath-Mistral-7B, oshizo/japanese-e5-mistral-7b_slerp, cognitivecomputations/dolphin-2.6-mistral-7b-dpo-laser and s3nh/Mistral-7B-Evol-Instruct-Chinese. Provider: macadeliccc. Access: Open.
Aggregate Rating 1417 ± 33, rank #5330 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - HellaSwag | 84.85 | Normalized accuracy (%) (10-shot) | 69.3 |
| Open LLM Leaderboard v1 - GSM8K | 56.63 | Accuracy (%) (5-shot) | 68.4 |
| Open LLM Leaderboard v1 - ARC Challenge | 63.74 | Normalized accuracy (%) (25-shot) | 64.1 |
| Open LLM Leaderboard v1 - WinoGrande | 78.45 | Accuracy (%) (5-shot) | 63.3 |
| Open LLM Leaderboard v1 - MMLU | 63.57 | Accuracy (%) (5-shot) | 61.6 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 53.78 | MC2 (%) (0-shot) | 60.7 |
Interactive version: theaggregate.ai/model?slug=polyglot-math-4x7b · How It Works · Data refreshed daily, snapshot 2026-11-08.