QwQ-32B-Preview

Overview

An experimental research model focused on advancing AI reasoning capabilities, particularly excelling in mathematics and programming. Features deep introspection and self-questioning abilities while having some limitations in language mixing and recursive reasoning patterns.

QwQ-32B-Preview was released on November 28, 2024. API access is available through 4 providers, including DeepInfra, Hyperbolic and others.

Benchmarks

QwQ-32B-Preview Performance Across Datasets

Scores sourced from the model's scorecard, paper, or official blog posts

llm-stats.com - Mon Dec 15 2025

Pricing

Pricing, performance, and capabilities for QwQ-32B-Preview across different providers:

Provider	Input ($/M)	Output ($/M)	Max Input	Max Output	Latency (s)	Throughput	Quantization	Input	Output
DeepInfra	$0.15	$0.60	32.8K	32.8K	0.44	76.04 tok/s	—	Text Image Audio Video	Text Image Audio Video
Hyperbolic	$0.20	$0.20	32.8K	32.8K	1.05	31.9 tok/s	—	Text Image Audio Video	Text Image Audio Video
Fireworks	$0.89	$0.89	32.8K	32.8K	0.53	99.15 tok/s	—	Text Image Audio Video	Text Image Audio Video
Together	$1.20	$1.20	32.8K	32.8K	0.74	62.14 tok/s	—	Text Image Audio Video	Text Image Audio Video

Price Comparison for QwQ-32B-Preview

Price per 1M input tokens (USD), lower is better

llm-stats.com - Mon Dec 15 2025

Throughput Comparison for QwQ-32B-Preview

Tokens per second, higher is better

llm-stats.com - Mon Dec 15 2025

Latency Comparison for QwQ-32B-Preview

Time to first token (s), lower is better

llm-stats.com - Mon Dec 15 2025

Example Outputs

API Access