Skip to content
Tokenly.
Demo fixture — not an offerDemo fixture

Qwen2.5 72B Instruct

Qwen/Qwen2.5-72B-Instruct

Alibaba's 72-billion-parameter open-weight model with strong Chinese and English performance, structured output support and a 128K-token context window.

Qwen2.5 72B Instruct
Model IDQwen/Qwen2.5-72B-Instruct
DeveloperAlibaba Qwen
Serving providerPending confirmation
Capabilitieschatreasoningfunction-callingstructured-outputlong-context
Modalitiestext
Context window131K tokens
Pricing
Input
Not yet published
Output
Not yet published
LimitationsText-only. Quantized or hosted variants may behave differently from the model card.
SourceModel card ↗
Last verified18 Nov 2026