Qwen3-8B-Base — benchmark results

Alibaba's open 8B dense pretrained base from the April 2025 Qwen3 family, trained on ~36T tokens across 119 languages. Provider: Alibaba. Released 2025-04-28. Access: Open.

Unified ELO 1527 ± 14, rank #692 of 1776 rated models, from 246 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Ukrainian LLM - Global MMLU Full UK Abstract Algebra61Score (%)100
Ukrainian LLM - Global MMLU Full UK Astronomy84.21Score (%)100
Ukrainian LLM - Global MMLU Full UK College Chemistry54Score (%)100
Ukrainian LLM - Global MMLU Full UK College Computer Science70Score (%)100
Ukrainian LLM - Global MMLU Full UK College Mathematics52Score (%)100
Ukrainian LLM - Global MMLU Full UK College Physics62.75Score (%)100
Ukrainian LLM - Global MMLU Full UK Conceptual Physics79.15Score (%)100
Ukrainian LLM - Global MMLU Full UK Econometrics63.16Score (%)100
Ukrainian LLM - Global MMLU Full UK Electrical Engineering75.86Score (%)100
Ukrainian LLM - Global MMLU Full UK Elementary Mathematics80.95Score (%)100
Ukrainian LLM - Global MMLU Full UK Formal Logic62.7Score (%)100
Ukrainian LLM - Global MMLU Full UK High School Biology86.45Score (%)100

Interactive version: theaggregate.ai/model?slug=qwen3-8b-base · How the rankings work · Data refreshed daily, snapshot 2026-07-22.