Hippolyta-7B-bf16: benchmark results
Kquant03's Mistral-7B-v0.1 fine-tune, whose card credits work with the ConvexAI team on reformatted public datasets and a small private component. Its published configuration is a dense 32-layer Mistral decoder; the separately linked GGUF conversion is not equated with this release. Provider: Kquant03. Access: Open.
Aggregate Rating 1366 ± 36, rank #7272 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 55.74 | MC2 (%) (0-shot) | 65.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 60.58 | Normalized accuracy (%) (25-shot) | 48.4 |
| Open LLM Leaderboard v1 - MMLU | 57.71 | Accuracy (%) (5-shot) | 41.7 |
| Open LLM Leaderboard v1 - HellaSwag | 79.98 | Normalized accuracy (%) (10-shot) | 39.7 |
| Open LLM Leaderboard v1 - WinoGrande | 73.95 | Accuracy (%) (5-shot) | 34.2 |
| Open LLM Leaderboard v1 - GSM8K | 1.82 | Accuracy (%) (5-shot) | 18.4 |
Interactive version: theaggregate.ai/model?slug=hippolyta-7b-bf16 · How It Works · Data refreshed daily, snapshot 2026-11-08.