Everyone-Coder-4x7B-Base: benchmark results
A community coding mixture assembled by rombodawg from four Mistral-family experts, using a Mixtral architecture with two experts selected per token. Provider: rombodawg. Access: Open.
Aggregate Rating 1397 ± 43, rank #6259 of 10916 rated models, from 13 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Portuguese LLM - FaQuAD NLI | 70.39 | Macro F1 (%) | 61 |
| Open Portuguese LLM - OAB Exams | 44.65 | Accuracy (%) | 58.8 |
| Open Portuguese LLM - ENEM | 64.1 | Accuracy (%) | 57.7 |
| Open Portuguese LLM - ASSIN2 RTE | 89.49 | Macro F1 (%) | 50.5 |
| Open Portuguese LLM - BLUEX | 51.6 | Accuracy (%) | 49.9 |
| Open Chinese LLM - GSM8K | 40.86 | Accuracy (%) | 40.1 |
| Open Chinese LLM - CMMLU | 48.58 | Accuracy (%) | 29.9 |
| Open Chinese LLM - HellaSwag | 55.64 | Accuracy (%) | 27.2 |
| Open Chinese LLM - TruthfulQA MC | 51.33 | Accuracy (%) | 27.2 |
| Open Chinese LLM Leaderboard | 51.32 | Official Nonzero-Task Mean (%) | 21.9 |
| Open Chinese LLM - ARC Challenge | 46.33 | Accuracy (%) | 21.7 |
| Open Chinese LLM - C-Eval Semantic | 56.05 | Accuracy (%) | 17.5 |
Interactive version: theaggregate.ai/model?slug=everyone-coder-4x7b-base · How It Works · Data refreshed daily, snapshot 2026-11-08.