Qwen3 30B A3B Instruct FP8: benchmark results

Provider: Alibaba. Released 2025-04-29.

Aggregate Rating 1593 ± 28, rank #1082 of 10916 rated models, from 28 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Medmarks - LongHealth Task 288.58Score (%)69.9
Medmarks - PubMedQA76.27Score (%)61.9
Medmarks - SCTpublic66.92Score (%)61.4
Medmarks - SuperGPQA Medicine Easy56.11Score (%)59.1
Medmarks - M-ARC43Score (%)58
Medmarks - MedQA86.65Score (%)56.8
Medmarks - CareQA EN88.67Score (%)55.7
Medmarks - MMLU-Pro Health74.74Score (%)55.7
Medmarks - MedMCQA71.53Score (%)55.7
Medmarks - MedCalc-Bench55.73Score (%)54.5
Medmarks - SuperGPQA Medicine Hard39.78Score (%)52.8
Medmarks - MedXpertQA Understanding24.67Score (%)50

Interactive version: theaggregate.ai/model?slug=qwen3-30b-a3b-instruct-fp8 · How It Works · Data refreshed daily, snapshot 2026-11-08.