Qwen 3.6 35B A3B — benchmark results

Alibaba's efficiency-focused open Qwen 3.6 35B A3B MoE (3B active). Provider: Alibaba. Released 2026-04-16. Access: Open.

Unified ELO 1658 ± 11, rank #350 of 1776 rated models, from 162 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
JudgeBench Math94.64Accuracy (%)97.1
LLM Stats (MMBench-V1.1)92.8Score (%)97.1
LLM Stats (Hallusion Bench)69.8Score (%)93.3
K-MetBench82.9Accuracy (self-reported)93.1
MWS Vision Bench69.4Overall Score (%)91.9
HardMTBench90.68HardMTBench zh-en GEMBA-DA (self-reported)89.5
AI for Education Pedagogy - Social studies86.36Accuracy (%)88.6
MMGist55.2Macro ↑ (self-reported)88.5
LLM Stats (CC-OCR)81.9Score (%)88.2
DiagnosticIQ69.9Macro % D.IQ (self-reported)87.5
LLM Stats (ODinW)50.8Score (%)86.7
RewardBench 2 Math90.16Accuracy (%)86.3

Interactive version: theaggregate.ai/model?slug=qwen-3-6-35b-a3b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.