Alibaba's Qwen takes on Kimi K3 with open-weight Qwen 3.8, says model is "second only to Fable 5"
Quick Answer
Alibaba has launched Qwen 3.8, an open-weight model with 2.4 trillion parameters, claiming it ranks just below Fable 5.
Quick Take
This model is expected to outperform Qwen 3.7-Max in coding and productivity tasks, with multimodal capabilities for processing various media. Open weights will be available soon at a 10% introductory price.
Key Points
- Qwen 3.8 features 2.4 trillion parameters, enhancing its performance in complex tasks.
- It is Alibaba's first capable of processing images, videos, and documents.
- The model is expected to disrupt Kimi K3's market presence and strategy.
- Open weights will be available soon through Alibaba's Token Plan at reduced prices.
- Moonshot AI, with $300 million in revenue, plans to go public within six months.
📖 Reader Mode
~1 min readAlibaba has unveiled Qwen 3.8, its latest open-weight model with 2.4 trillion parameters. The Qwen team says it matches frontier models and trails only Fable 5. A preview is available through Alibaba's Token Plan, Qoder, and QoderWork at 10 percent of the standard price. Open weights are coming "soon."
Alibaba says Qwen 3.8 should outperform Qwen 3.7-Max, especially in coding and complex productivity tasks such as full-stack development, data analysis, and office workflows. Developer Shuai Bai says it's also the team's first multimodal model with more than 1 trillion parameters and can process images, videos, and documents. No benchmark results are available yet.
Qwen's push is likely partly aimed at disrupting Kimi K3's momentum. Moonshot AI has promised open weights, but access currently runs through its chat app and API, turning interest into customers. Qwen's release could disrupt that strategy. After reaching $300 million in annual recurring revenue in June, Moonshot plans to go public in as little as six months, Bloomberg reports.
— Originally published at the-decoder.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from The Decoder
See more →
An AI model programmed nonstop for 19 days on a single MirrorCode task that cost $2,600 to run
Epoch AI's MirrorCode benchmark reveals Claude Opus 4.7 as the leader with a 56% solve rate, reconstructing a 16,000-line toolkit in 14 hours. Despite this, all models tested struggle with the most complex tasks, highlighting limitations in current AI capabilities. The single task consumed $2,600 over 19 days, raising questions about cost-effectiveness in AI development.
