Qwen 3.6 Plus

Qwen 3.6 Plus, released April 2 2026, quadrupled the context window to 1M tokens — matching Claude’s context at a fraction of the cost.

Key specs

Spec	Value
Context window	1M tokens (native, 4x Qwen 3.5)
Architecture	MoE with dense variants
Pricing	$0.30/$1.20 per M tokens
Open weights	Yes

Qwen 3.x family

The Qwen 3 series offers dense and MoE models across multiple sizes, targeting everything from edge deployment to frontier reasoning.

Strengths

• 1M native context at open-source pricing
• Competitive benchmarks for its price tier
• Strong multilingual performance (especially CJK languages)
• Active release cadence — rapid iteration

Weaknesses

• Benchmark scores trail Western frontier models by 10+ points
• Less ecosystem support than Llama in Western markets
• Documentation primarily in Chinese