Qwen3 4B

alibaba

This qwen3 hybrid reasoning model enables seamless switching between thinking and non-thinking modes during conversations, achieving SOTA reasoning performance at its scale. It shows significant improvements in human preference alignment, creative writing, role-playing, multi-turn dialogue, and instruction following, delivering a greatly enhanced user experience.

Try Now

Capabilities

Tool Use

Extended Thinking

Example Use Cases

Compact model with surprisingly strong reasoning

Local deployment with thinking capability

Small model for creative writing or dialogue

Technical Specifications

Context Window

131,072 tokens

Max Output

8,192 tokens

Cache Miss Cost

$0.11 per 1M tokens

Non-Reasoning Cost

$0.42 per 1M tokens

Reasoning Cost

$1.26 per 1M tokens

Web Search Cost

$15 per 1K calls

Code Execution Cost

$0.19 per 1K calls

⚠️ Legacy

Made legacy on

Reason

Untested

Recommended Replacement

Qwen3 Max