pythia-2.8B-deduped: benchmark results

Provider: EleutherAI. Released 2023-04-03. Access: Open.

Aggregate Rating 1135 ± 15, rank #10218 of 10916 rated models, from 84 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MMLU-by-task - High School Statistics44.91Accuracy (%)73.4
MMLU-by-task - High School Physics31.79Accuracy (%)61.3
MMLU-by-task - Global Facts33Accuracy (%)51.6
MMLU-by-task - College Chemistry30Accuracy (%)40.7
MMLU-by-task - High School Mathematics26.3Accuracy (%)40
MMLU-by-task - Professional Medicine43.38Accuracy (%)38.7
MMLU-by-task - College Mathematics30Accuracy (%)38.1
MMLU-by-task - Abstract Algebra27Accuracy (%)34
MMLU-by-task - Public Relations35.45Accuracy (%)33
MMLU-by-task - Elementary Mathematics26.46Accuracy (%)32.7
MMLU-by-task - High School European History28.48Accuracy (%)30.2
MMLU-by-task - High School Macroeconomics32.05Accuracy (%)29.4

Interactive version: theaggregate.ai/model?slug=pythia-2-8b-deduped · How It Works · Data refreshed daily, snapshot 2026-11-08.