Yi 1.5 34B — benchmark results

01.AI's open 34B upgrade of Yi, continually pre-trained on 500B extra tokens for stronger coding, math, reasoning, and instruction following. Provider: 01.AI. Released 2024-05-13. Access: Open.

Unified ELO 1499 ± 11, rank #801 of 1776 rated models, from 126 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval English Common Sense Reasoning94.53Common Sense Reasoning Average Score (%)99.4
EuroEval Spanish NLU - MLQA ES69.63Reading comprehension Score (%)97.8
EuroEval English77.14Average Score (%)97
EuroEval English NLU - SQuAD88.19Reading comprehension Score (%)96.6
EuroEval French NLU - FQuAD75.71Reading comprehension Score (%)96.1
EuroEval Finnish NLU - Tydiqa FI75.98Reading comprehension Score (%)95.7
EuroEval German NLU - Germanquad68.22Reading comprehension Score (%)95.3
EuroEval Portuguese NLU - MultiWikiQA PT77.91Reading comprehension Score (%)95.2
EuroEval Dutch NLU - SQuAD NL81.65Reading comprehension Score (%)95.1
EuroEval Italian NLU - SQuAD IT74.69Reading comprehension Score (%)95.1
EuroEval Swedish NLU - Swerec79.83Sentiment classification Score (%)95.1
EuroEval English NLU - SST-569.38Sentiment classification Score (%)94.5

Interactive version: theaggregate.ai/model?slug=yi-1-5-34b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.