NeuralSynthesis-7B-v0.3: benchmark results
Kukedlc's model-stock merge of six Mistral-7B variants (including YamshadowExperiment28 and M7-7b), built with LazyMergekit. Provider: Other. Released 2024-04-07. Access: Open.
Unified ELO 1482 ± 1, rank #777 of 1392 rated models, from 14 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Chinese LLM - TruthfulQA MC | 63.76 | Accuracy (%) | 94.7 |
| Open LLM Leaderboard - MuSR | 13.39 | Score | 73.8 |
| Open Chinese LLM - ARC Challenge | 54.01 | Accuracy (%) | 61.9 |
| Open Chinese LLM - HellaSwag | 60.88 | Accuracy (%) | 59.6 |
| Open LLM Leaderboard - BBH | 31.81 | Score | 58.6 |
| Open Chinese LLM - WinoGrande | 63.46 | Accuracy (%) | 57.3 |
| Open Chinese LLM Leaderboard | 56.96 | Average Score (%) | 57 |
| Open Chinese LLM - GSM8K | 47.23 | Accuracy (%) | 56.1 |
| Open LLM Leaderboard - IFEval | 40.78 | Score | 41.1 |
| Open LLM Leaderboard - MATH Level 5 | 7.78 | Score | 40.1 |
| Open LLM Leaderboard - MMLU-Pro | 22.78 | Score | 38.1 |
| Open LLM Leaderboard - GPQA | 4.03 | Score | 35.2 |
Interactive version: theaggregate.ai/model?slug=neuralsynthesis-7b-v0-3 · How It Works · Data refreshed daily, snapshot 2026-09-05.