Liph.42: benchmark results
Liph.42 is published by liminerity with a 32-layer Phi causal-language-model configuration. Its card does not provide a reproducible training recipe or exact training-data disclosure. Provider: liminerity. Access: Open.
Aggregate Rating 1383 ± 34, rank #6767 of 10916 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 59.82 | Accuracy (%) (5-shot) | 71.7 |
| Open LLM Leaderboard v1 - ARC Challenge | 62.29 | Normalized accuracy (%) (25-shot) | 57.1 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 52.08 | MC2 (%) (0-shot) | 54.5 |
| Open LLM Leaderboard v1 - MMLU | 58.2 | Accuracy (%) (5-shot) | 42.8 |
| Open LLM Leaderboard v1 - WinoGrande | 74.98 | Accuracy (%) (5-shot) | 39.7 |
| Open LLM Leaderboard v1 - HellaSwag | 77.12 | Normalized accuracy (%) (10-shot) | 30.2 |
Interactive version: theaggregate.ai/model?slug=liph-42 · How It Works · Data refreshed daily, snapshot 2026-11-08.