MODEL FAMILYLANGUAGEfamilyAlibaba Cloud

Qwen

Open-source language model project

Alibaba Cloud's Qwen is a Transformer-based multilingual language model project/series. The official introductory blog says Qwen is the base language model in the project, distinguishes Qwen from Qwen-Chat, and lists released model sizes including Qwen-1.8B, Qwen-7B, Qwen-14B, and Qwen-72B.

Pricing

Artificial Analysis · Mar 30, 2026, 1:01 PM

Input / 1M

$1.20

Output / 1M

$6.00

Blended / 1M

$2.40

Output TPS

35.27

TTFT (s)

1.63

Model Intelligence

Context window

32,768 tokens

Arena ranking

Benchmarkable

Model level

family

Intelligence Index

39.9

Coding Index

30.5

GPQA

0.86

HLE

0.26

SciCode

0.43

IFBench

0.71

LCR

0.66

TerminalBench Hard

0.24

TAU2

0.84

Recent stories

1 linked story

newsSECONDARY2026-03-28

ATLAS benchmarks Qwen3-14B at 74.6% LiveCodeBench on one RTX 5060 Ti

The ATLAS harness says a frozen Qwen3-14B Q4 model on one RTX 5060 Ti reached 74.6% pass@1-v(k=3) on LiveCodeBench v5 through multi-pass repair and selection. The result shifts comparison toward harness design, though HN commenters note it is not a one-shot head-to-head with hosted frontier models.