ruGPT-3 (Medium): benchmark results

Provider: Other. Access: Open.

Unified ELO 1263 ± 1, rank #1906 of 1935 rated models, from 31 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
MERA Original Paper - ruOpenBookQA27.3Accuracy (%; 5-shot, log-likelihood)72.2
MERA Original Paper - ruMMLU27.1Accuracy (%; 5-shot, log-likelihood)66.7
MERA Original Paper20.1Total Score (%; mean of 17 problem-solving and exam tasks, M63.9
MERA Original Paper - SimpleAr0.8Exact Match (%; 5-shot, greedy generation)61.1
MERA Original Paper - ruModAr0.1Exact Match (%; greedy generation)58.3
MERA Original Paper - ruMultiAr1.2Exact Match (%; 5-shot, greedy generation)58.3
MERA Original Paper - ruWorldTree25.1Accuracy (%; 5-shot, log-likelihood)47.2
MERA Original Paper - MathLogicQA24.8Accuracy (%; 5-shot, log-likelihood)22.2
MERA - CheGeKa1.56F1 (%)16.7
MERA - USE6.96Grade, normalized (%)15.1
MERA - LCS7.4Accuracy (%)11.2
MERA - ruHumanEval0.12pass@1 (%)9.6

Interactive version: theaggregate.ai/model?slug=rugpt-3-medium · How It Works · Data refreshed daily, snapshot 2026-09-25.