Qwen 3.5 0.8B (Thinking) — benchmark results

Provider: Alibaba. Released 2026-03-02. Access: Open.

Unified ELO 1261 ± 63, rank #1575 of 1756 rated models, from 37 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience-35.18Score58.9
MedLayXPlain58.2S (self-reported)51.6
AA TAU-2 Bench47.66Accuracy (%)51.4
CritPt0Accuracy (self-reported)30.3
AA CritPt0Accuracy (%)27.2
Artificial Analysis Intelligence Index5.46Intelligence Index16.9
AA Long Context Reasoning5.33Accuracy (%)16.8
AA Omniscience - Software Engineering (SWE) - Julia0Accuracy (%)8
GDPval-AA125Elo6.8
AA Terminal-Bench Hard0Accuracy (%)5.5
AA Omniscience - Software Engineering (SWE) - Dart4Accuracy (%)3.4
AA IFBench21.5Accuracy (%)3.1

Interactive version: aibenchmarks.dev/model?slug=qwen-3-5-0-8b-thinking · How the rankings work · Data refreshed daily, snapshot 2026-07-17.