Archive · 20 Jul 2026
Alibaba Launches Qwen 3.8 AI Model
Alibaba has introduced a new AI model called Qwen 3.8, which has 2.4 trillion parameters.
This model is very powerful and performs almost as well as the best AI models available today.
It is better than the previous version, Qwen 3.7 Max, especially in tasks like coding, reasoning, and doing complex work.
You can try it out through Alibaba Token, Qoder, and Qoder Work, and it costs much less than usual.
Alibaba is also trying to compete with another company, Moonshot AI, which has a model called Kimi K3.
Moonshot AI says their model is the largest open-source AI model and performs very well, but these claims have not been checked by independent experts.
Both companies are showing that Chinese AI labs are now making very advanced AI models.
Alibaba has launched Qwen 3.8, an AI model with 2.4 trillion parameters.
Qwen 3.8 performs on par with leading frontier AI models, trailing only Anthropic’s Claude Fable 5 in select benchmarks.
The model is available for preview through Alibaba Token, Qoder, and Qoder Work, with the latter two priced at 10% of the standard cost.
Qwen 3.8 significantly outperforms Qwen 3.7 Max in coding, reasoning, and complex productivity tasks.
The launch is aimed at competing with Moonshot AI’s Kimi K3, which claims state-of-the-art performance but lacks independent verification.
- Who
- Alibaba
- What
- Launched Qwen 3.8, an AI model with 2.4 trillion parameters
- When
- At the World Artificial Intelligence Conference in Shanghai
- Where
- Shanghai, China
- Why
- To compete with leading AI models and disrupt Moonshot AI’s Kimi K3
Alibaba
Moonshot AI
Model Performance
Alibaba
Qwen 3.8 is one of the most powerful models, surpassed only by Anthropic's Claude Fable 5.
Moonshot AI
Kimi K3 is the world's largest open-source AI model with 2.8 trillion parameters and claims state-of-the-art performance.
Benchmarking Claims
Alibaba
Alibaba claims Qwen 3.8 outperforms Qwen 3.7 Max in coding, reasoning, and complex productivity tasks.
Moonshot AI
Moonshot AI claims Kimi K3 outperforms Claude Fable 5 and GPT-5.6 Sol in software project building and long-term software engineering tasks.
- Model Name
- Qwen 3.8
- Parameters
- 2.4 trillion
- Performance Comparison
- On par with leading frontier AI models, trailing only Anthropic’s Claude Fable 5 in select benchmarks
- Availability
- Preview available through Alibaba Token, Qoder, and Qoder Work
- Cost
- Qoder and Qoder Work priced at 10% of standard cost
- Improvements Over Qwen 3.7 Max
- Significant performance boost in coding, reasoning, and complex productivity tasks
- Competitor
- Moonshot AI’s Kimi K3
- Kimi K3 Parameters
- 2.8 trillion
- Kimi K3 Performance Claims
- State-of-the-art performance across a wide range of benchmarks
- Independent Verification
- Neither Alibaba's nor Moonshot AI's claims have been independently verified
- Chinese AI Sector
- Chinese AI labs have entered the era of trillion-parameter models
- Political Signal
- Chinese President Xi Jinping's commitment to fostering open source and collaboration in AI
Sources
Alibaba launches Qwen 3.8 with 24 trillion parameters, targets frontier AI models
After Kimi K3, Alibaba unveils Qwen 3.8, claims it's second only to Claude Fable 5
Following the Kimi K3, Alibaba launches a major AI innovation: Qwen 3.8, which the company claims is surpassed only by Anthropic's Claude Fable 5
