oasst-pythia-12B-6000-steps: benchmark results

Provider: Other. Access: Open.

Aggregate Rating 1140 ± 17, rank #10110 of 10916 rated models, from 68 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MMLU-by-task - College Physics28.43Accuracy (%)67.4
MMLU-by-task - Abstract Algebra30Accuracy (%)52.2
MMLU-by-task - Global Facts32Accuracy (%)44.3
MMLU-by-task - Elementary Mathematics27.25Accuracy (%)41.4
MMLU-by-task - High School Mathematics26.3Accuracy (%)40
MMLU-by-task - ARC Challenge41.47Accuracy (%)36.3
MMLU-by-task - HellaSwag52.4Accuracy (%)34.8
MMLU-by-task - TruthfulQA MC125.34Accuracy (%)34.6
MMLU-by-task - High School Computer Science33Accuracy (%)33.8
MMLU-by-task - High School Chemistry28.57Accuracy (%)33.2
MMLU-by-task - Astronomy30.92Accuracy (%)32.4
MMLU-by-task - High School US History28.92Accuracy (%)32

Interactive version: theaggregate.ai/model?slug=oasst-pythia-12b-6000-steps · How It Works · Data refreshed daily, snapshot 2026-11-08.