Keimodel
Log inStart free
Qwen

Qwen

Alibaba's multilingual frontier models

Qwen is Alibaba Cloud's AI model family, built to serve global users with particular strength in Chinese and other Asian languages. The Qwen 3 generation uses mixture-of-experts architecture and delivers impressive benchmark results, especially on coding and mathematical reasoning. Qwen models are released as open weights and available cheaply via API.

Founded 1999·7 models·Visit Qwen

7 Models

Qwen3.8 Max

Alibaba's flagship Qwen

3 cr
Qwen3.8 Max is the flagship model in Alibaba's Qwen3.8 series, the general-availability successor to the Qwen3.8 Max Preview. It is a multimodal reasoning model intended for complex reasoning, visual understanding,...
1M ctxVisionTool use

Qwen3.8 27B

Laptop-class Qwen 3.8

2 cr
Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...
262k ctxVisionTool use

Qwen3.7 Flash

Among the cheapest capable models

2 cr
Qwen3.7 Flash is a vision-language reasoning model from Alibaba. It is suited for multimodal agents, visual coding, search, and computer interaction, with strengths in object recognition, spatial understanding, and real-world...
1M ctxVisionTool use

Qwen3 235B

Alibaba's flagship 235B MoE model

5 cr
Qwen3 235B is Alibaba's most capable model, using a mixture-of-experts architecture with 235 billion total parameters and 22 billion active. It delivers top-tier performance on multilingual reasoning, coding, and math tasks.
128k ctxTool useJSON mode

Qwen3 32B

Dense 32B model with hybrid thinking

3 cr
Qwen3 32B is a dense model that supports both standard and thinking (chain-of-thought) modes, allowing users to trade latency for reasoning depth depending on the task at hand.
128k ctxTool useJSON mode

Qwen3 VL 235B

Frontier open-source vision-language model

5 cr
Qwen3 VL 235B is Alibaba's most capable vision-language model, combining the 235B MoE architecture with advanced multimodal understanding. It sets new records for open-weight models on visual reasoning, document analysis, and image comprehension benchmarks.
128k ctxVisionTool use

Qwen2.5 VL 72B

Leading open-source vision-language model

3 cr
Qwen2.5 VL 72B is one of the most capable open-source vision-language models, able to analyse images, charts, documents, and screenshots with strong accuracy. It outperforms GPT-4V on several vision benchmarks.
128k ctxVisionTool use

Other providers

OpenAIOpenAIAnthropicAnthropicGoogleGoogleMetaMeta