QwenQwen 3Open SourceArena #260Mar 5, 2025

Qwen: QwQ 32B

QwQ is the reasoning model of the Qwen series. Compared with conventional instruction-tuned models, QwQ, which is capable of thinking and reasoning, can achieve significantly enhanced performance in downstream tasks, especially hard problems. QwQ-32B is the medium-sized reasoning model, which is capable of achieving competitive performance against state-of-the-art reasoning models, e.g., DeepSeek-R1, o1-mini.

Context Window
33K
tokens
Max Output
33K
tokens
Released
Mar 5, 2025
Arena Rank
#260
of 305 models
Output Speed
125
tokens/sec
Time to First Token
3.2s
TTFT

Capabilities

👁Vision
🧠Reasoning
🔧Tool Calling
Prompt Caching
🖥Computer Use
🎨Image Generation

Supported Parameters

Frequency Penalty
Reduce repetition
Include Reasoning
Show reasoning tokens
Logit Bias
Adjust token weights
Max Tokens
Output length limit
min_p
Presence Penalty
Encourage new topics
Reasoning
Extended thinking
Repetition Penalty
Penalize repeated tokens
Response Format
JSON mode / structured output
Seed
Deterministic outputs
Stop Sequences
Custom stop tokens
structured_outputs
Temperature
Controls randomness
Tool Choice
Control tool usage
Tool Calling
Function calling support
Top K
Top-K sampling
Top P
Nucleus sampling

Pricing Comparison

RouterInput / 1MOutput / 1MCached Input / 1M
OpenRouter$0.15$0.40
Martian$0.15$0.40

Benchmarks

Open LLM Leaderboard
AverageOpen LLM Leaderboard
34.12/100
IFEvalOpen LLM Leaderboard
40.35/100
BBHOpen LLM Leaderboard
53.39/100
MATH Lvl 5Open LLM Leaderboard
44.94/100
GPQAOpen LLM Leaderboard
4.25/100
MUSROpen LLM Leaderboard
9.81/100
MMLU-PROOpen LLM Leaderboard
51.98/100
Artificial Analysis
Intelligence IndexArtificial Analysis
56/100
Coding IndexArtificial Analysis
50/100
Math IndexArtificial Analysis
82/100
MMLU-PROArtificial Analysis
0.738/1
GPQA DiamondArtificial Analysis
0.618/1
MATH-500Artificial Analysis
0.928/1
AIME 2024Artificial Analysis
0.55/1
LiveCodeBenchArtificial Analysis
0.518/1
Aider Polyglot
Aider PolyglotAider Polyglot
8/100

Model IDs

OpenRouterqwen/qwq-32b
Hugging FaceQwen/QwQ-32B

Tags

reasoningtool-calling
Compare with another model

Related Models