Qwen 3 235B A22B (Thinking): benchmark results

Alibaba Qwen 3 235B A22B evaluated with thinking enabled. Provider: Alibaba. Released 2025-04-29. Access: Open.

Aggregate Rating 1617 ± 7, rank #654 of 2168 rated models, from 279 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
DiagFlowBench - Termination Recognition98.5Termination recognition (%): share of terminal-node turns at100
EmoHarbor4.13Mean of 10 Rubric Dimensions (1-5 Likert)100
OpenCompass LLM - Code - English (CompassBench 2504)61.5Score (%)100
OpenCompass LLM - Reasoning (CompassBench 2504)92.4Score (%)100
OpenCompass LLM - Reasoning - English (CompassBench 2504)92.6Score (%)100
OpenCompass Reasoning - Academic - English (CompassBench 2504)94.6Score (%)100
OpenCompass Reasoning - Common - Chinese (CompassBench 2504)92.4Score (%)100
OpenCompass Reasoning - Puzzle - English (CompassBench 2504)96Score (%)100
PodBench - Thinking Mode - Instruction Following94.32Instruction Following (0-100; judge-derived checklist, items100
TeleCom-Bench - Root Cause Diagnosis71.49Exact match (%; mean of a fault-agent set and a single-domai100
WuYuEval - Expert Module - Judge Score0.87Mean anchor-calibrated LLM-as-a-judge score (0-1)100
OpenCompass Code - Competition - English (CompassBench 2504)78.7Score (%)98.1

Interactive version: theaggregate.ai/model?slug=qwen-3-235b-a22b-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.