U-Amethyst-20B: benchmark results

Provider: Undi95. Access: Open.

Aggregate Rating 1362 ± 37, rank #7377 of 10916 rated models, from 6 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard v1 - TruthfulQA MC253.2MC2 (%) (0-shot)58.8
Open LLM Leaderboard v1 - ARC Challenge62.2Normalized accuracy (%) (25-shot)56.6
Open LLM Leaderboard v1 - HellaSwag83.11Normalized accuracy (%) (10-shot)56
Open LLM Leaderboard v1 - MMLU55.88Accuracy (%) (5-shot)37.6
Open LLM Leaderboard v1 - WinoGrande74.19Accuracy (%) (5-shot)35.4
Open LLM Leaderboard v1 - GSM8K5.31Accuracy (%) (5-shot)23.9

Interactive version: theaggregate.ai/model?slug=u-amethyst-20b · How It Works · Data refreshed daily, snapshot 2026-11-08.