Alibaba Cloud

16 models from Alibaba Cloud ranked on LMRank.

Alibaba Cloud models
Rank Model Tags Context Price (in / out)
19 Qwen3.7 Max
Intelligent Reasoning Multimodal
1,000,000 $2.50 / $7.50
22 Qwen3.6 Max Preview
Intelligent Reasoning Multimodal
262,144 $1.04 / $6.24
26 Qwen: Qwen3 235B A22B
Open Weight Intelligent Coding Reasoning
131,072 $0.45 / $1.82
42 Qwen: Qwen3.7 Plus
Intelligent Image Input Long Context
1,000,000 $0.40 / $1.60
43 Qwen3.6 35B A3B
Coding
262,144 $0.15 / $1.00
68 Qwen3.6 Plus
Intelligent Coding Long Context
1,000,000 $0.33 / $1.95
76 Qwen3.6 27B
Intelligent Coding Open Weight
262,144 $0.32 / $3.20
82 Qwen 3 72B
Intelligent Open Weight
131,072 $0.50 / $1.50
91 Qwen3.5 397B A17B
Intelligent Open Weight Multimodal
262,144 $0.39 / $2.34
100 Qwen3.5-27B
Open Weight
262,144 $0.20 / $1.56
103 Qwen3.5-122B-A10B
Open Weight
262,144 $0.26 / $2.08
107 Qwen3 Max Thinking
Reasoning Intelligent
262,144 $0.78 / $3.90
109 Qwen3.5-35B-A3B
Open Weight Fast
262,144 $0.14 / $1.00
113 Qwen3.5-9B
Open Weight Fast
262,144 $0.10 / $0.15
115 Qwen 2.5 72B
Intelligent Open Weight
32,768 $0.35 / $1.20
117 Qwen3 Coder Next
Coding Open Weight Fast
262,144 $0.11 / $0.80

In Depth: Alibaba Cloud

About Alibaba Cloud

Alibaba Cloud’s Qwen family delivers 6 active models, with top performer Qwen3.7 Max scoring 9.0/10 and ranking #13 on LMRank. Qwen offers developers open weights under Apache 2.0 and strong Chinese-English bilingual performance.

Alibaba Cloud follows an annual major version release (Qwen, Qwen2, Qwen2.5) with quarterly updates and domain-specific variants. Most models are open-source under Apache 2.0, with weights on Hugging Face and ModelScope. API access is available through Alibaba Cloud’s Tongyi Qianwen service.

Model Lineup

Qwen3.7 Max leads the lab at rank #13 with a 9.0/10 score. The lineup also includes prior-generation models like Qwen2.5-72B, Qwen2.5-32B, 14B, 7B, and smaller variants for diverse deployment needs.

The Qwen2.5 series spans flagship 72B to edge-ready 0.5B models, including specialized Coder and Math variants. Qwen3.7 Max is the top performer. Each tier targets specific compute budgets: large models for cloud servers, smaller ones for edge devices.

Pricing & Access

Qwen2.5-72B API pricing is approximately $2 per million input tokens and $8 per million output tokens. Smaller sizes cost less. Model weights are freely available under Apache 2.0 license, enabling self-hosting with sufficient GPU resources.

Compared to open-source peers like LLaMA 3 and Mistral, Qwen’s API pricing is competitive but not the cheapest. Enterprise licenses are available. Self-hosting the 72B model demands significant GPU investment, making API access a practical entry point for many teams.

Best For & When to Look Elsewhere

Qwen models excel in Chinese-English bilingual tasks, coding, and math reasoning. They offer permissive licensing for customization. Consider alternatives if you need minimal compute or stronger Western-language community support.

  • Bilingual Chinese-English chatbot development
  • Complex math and code reasoning tasks
  • Custom fine-tuning with open weights
  • Cloud-native deployment via Alibaba ecosystem