Experiment22-7B: benchmark results
A 32-layer Mistral-family checkpoint released as Experiment22-7B. The exact evaluated configuration uses 32,000 vocabulary entries; the release does not document a reproducible training recipe or establish equivalence with other numbered experiments. Provider: Other. Access: Open.
Aggregate Rating 1460 ± 42, rank #3353 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 79.47 | MC2 (%) (0-shot) | 99.7 |
| Open LLM Leaderboard v1 - WinoGrande | 84.77 | Accuracy (%) (5-shot) | 95 |
| Open LLM Leaderboard v1 - HellaSwag | 88.89 | Normalized accuracy (%) (10-shot) | 94.8 |
| Open LLM Leaderboard v1 - ARC Challenge | 71.5 | Normalized accuracy (%) (25-shot) | 89.4 |
| Open LLM Leaderboard v1 - GSM8K | 66.64 | Accuracy (%) (5-shot) | 82.1 |
| Open LLM Leaderboard v1 - MMLU | 64.13 | Accuracy (%) (5-shot) | 66 |
Interactive version: theaggregate.ai/model?slug=experiment22-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.