oasst-pythia-12B-reference: benchmark results

Provider: Other. Access: Open.

Aggregate Rating 1153 ± 16, rank #9916 of 10916 rated models, from 68 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MMLU-by-task - College Physics30.39Accuracy (%)74.7
MMLU-by-task - High School Physics32.45Accuracy (%)66
MMLU-by-task - Professional Medicine46.32Accuracy (%)50.6
MMLU-by-task - College Chemistry30Accuracy (%)40.7
MMLU-by-task - Abstract Algebra28Accuracy (%)39.6
MMLU-by-task - Elementary Mathematics26.98Accuracy (%)39.2
MMLU-by-task - Professional Accounting29.43Accuracy (%)37.6
MMLU-by-task - High School Statistics32.87Accuracy (%)37.3
MMLU-by-task - Security Studies37.55Accuracy (%)35.4
MMLU-by-task - Business Ethics36Accuracy (%)35.2
MMLU-by-task - ARC Challenge40.44Accuracy (%)34.5
MMLU-by-task - Clinical Knowledge33.21Accuracy (%)34.4

Interactive version: theaggregate.ai/model?slug=oasst-pythia-12b-reference · How It Works · Data refreshed daily, snapshot 2026-11-08.