Qwen 3.8 27B (xHigh): benchmark results

Provider: Alibaba. Released 2026-08-14. Access: Open.

Unified ELO 1668 ± 1, rank #187 of 1761 rated models, from 48 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Tau3 Banking48.04Success Rate (%)97.5
AA Long Context Reasoning82Accuracy (%)94.9
Vals AI SAGE52.4Accuracy (%)93.7
Artificial Analysis Intelligence Index41.65Intelligence Index91.6
AA GPQA Diamond90.51Accuracy (%)90.4
Vals AI Harvey Legal Agent Bench11.25Accuracy (%)90.4
AA-LCR77.33Accuracy (self-reported)88.4
AA Humanity's Last Exam33.92Accuracy (%)85.9
AA GDPval1467.93ELO84
Vals AI SWE-bench Verified86Resolved (%)83.9
AA CritPt5.43Accuracy (%)79.5
Vals AI GPQA88.89Accuracy (%)77

Interactive version: theaggregate.ai/model?slug=qwen-3-8-27b-xhigh · How It Works · Data refreshed daily, snapshot 2026-09-05.