QwQ-32B is a 32-billion-parameter open-weight reasoning large language model in the Qwen series, built on Qwen2.5-32B and enhanced with reinforcement learning for math, coding, general problem solving, instruction following, and agent/tool-use capabilities.
Pricing
Model profile · Current snapshot
Input / 1M
$0.66
Output / 1M
$1.00
Blended / 1M
$0.745
Output TPS
0
TTFT (s)
0
Model Intelligence
Context window
131,072 tokens
Arena ranking
13
Benchmarkable
Yes
Model level
release
Intelligence Index
13.4
Math Index
29
MMLU Pro
0.76
GPQA
0.59
HLE
0.07
LiveCodeBench
0.63
SciCode
0.36
MATH-500
0.96
AIME
0.78
AIME 2025
0.29
IFBench
0.39
LCR
0.26
Recent stories
0 linked stories
No linked stories yet.