StopCarbon-10.7B-v5: benchmark results
kekmodel's StopCarbon 10.7B v5, a mergekit SLERP merge of kyujinpy's Sakura-SOLAR-Instruct and jeonsworld's CarbonVillain-en-10.7B-v1. Provider: kekmodel. Released 2024-03-01. Access: Open.
Aggregate Rating 1441 ± 14, rank #1044 of 1656 rated models, from 46 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 71.84 | MC2 (%) (0-shot) | 92.3 |
| Open LLM Leaderboard v1 - HellaSwag | 88.48 | Normalized accuracy (%) (10-shot) | 92.2 |
| Open Korean LLM - KorNAT-Harmless | 67.82 | Normalized accuracy (%) | 90.7 |
| Open LLM Leaderboard v1 - WinoGrande | 83.58 | Accuracy (%) (5-shot) | 89.7 |
| Open LLM Leaderboard v1 - ARC Challenge | 70.99 | Normalized accuracy (%) (25-shot) | 87.5 |
| Open LLM Leaderboard v1 - MMLU | 66.34 | Accuracy (%) (5-shot) | 86.9 |
| Open Chinese LLM - TruthfulQA MC | 61.04 | Accuracy (%) | 85.8 |
| Open Portuguese LLM - ASSIN2 RTE | 92.93 | Macro F1 (%) | 85 |
| Open Korean LLM - Ko-EQ-Bench | 35.04 | EQ-Bench score | 84.8 |
| Open Korean LLM - Ko-GSM8K (flexible extract) | 51.1 | Exact match, flexible extract (%) | 84.5 |
| Open Korean LLM - KorNAT-Helpful | 49.85 | Normalized accuracy (%) | 82.1 |
| Open LLM Leaderboard v1 - GSM8K | 65.2 | Accuracy (%) (5-shot) | 79.5 |
Interactive version: theaggregate.ai/model?slug=stopcarbon-10-7b-v5 · How It Works · Data refreshed daily, snapshot 2026-10-20.