Alibaba Challenges Anthropic with its AI Qwen3.8-Max
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
After making waves with Moonshot, Alibaba is back in the spotlight with a new artificial intelligence model. Named Qwen3.8-Max, this model aims to be a serious competitor to the most advanced American solutions.
China, as the world's second-largest economic power, is increasingly demonstrating its ability to compete with the United States in the field of artificial intelligence. Chinese tech companies, such as Alibaba, continue to prove their potential by developing products that can measure up to international standards. Following the notable launch of Kimi K3 by Moonshot, Alibaba is presenting its Qwen3.8-Max model this week.
Qwen3.8-Max: A Credible Alternative to AI from Anthropic?
Alibaba recently unveiled its artificial intelligence model, Qwen3.8-Max, which it describes as a "new benchmark for coding and coworking." Equipped with 2.4 trillion parameters, this model reportedly shows performance comparable to Gemini 3.1-Pro and GPT-5.6 Sol, according to Alibaba's internal tests. It even approaches Claude Fable 5 from Anthropic in complex tasks such as multimodal reasoning, document analysis, and coding.
An interesting point to note is that this model will be accessible starting next week.
AI tools dedicated to coding have been crucial to Anthropic's success. Alibaba intends to stand out in this area as well, with a project titled "oh-my-cli," which Qwen3.8-Max developed "from scratch" during an autonomous coding session lasting over 10 days. This project is now available on GitHub.
For those interested, Alibaba announced that "the model weights will be made available as open source on Hugging Face and ModelScope starting next week." Additionally, the model will be exclusively available this week on QwenCloud.
In terms of costs, Qwen3.8-Max is particularly competitive compared to its rivals. Here’s a breakdown of the costs per million tokens:
- Qwen3.8-Max: $2 for input, $6 for output
- Claude Fable 5: $10 for input, $50 for output
- Kimi K3: $3 for input, $15 for output
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.