Alibaba выпустила Qwen3.8-Max — крупнейшую модель Китая, соперницу Fable 5 и GPT

Alibaba opened public access to Qwen3.8-Max on August 11, 2026 — its largest language model to date. The Chinese tech giant claims that the newcomer's performance closely approaches the level of Anthropic's Fable 5 and OpenAI's flagship systems, while among Chinese competitors, Qwen3.8-Max, according to the company, surpasses Kimi K3 from Moonshot AI.

What Qwen3.8-Max Is

Qwen3.8-Max is the largest model in the history of the Qwen lineup. Alibaba began introducing partners to the model as early as July 2026: at that time, in the Qwen blog, the company stated that the model "trails only Fable 5" — Anthropic's flagship. The public release took place on August 11, 2026 in open-weight format: the model's weights are open for free downloading, fine-tuning, and deployment on any infrastructure.

Key facts according to Alibaba:

• Public release — August 11, 2026
• Format — open weights (open-weight)
• Performance comparable to Fable 5 (Anthropic) and OpenAI's top systems
• Surpasses Kimi K3 from China's Moonshot AI
• Largest model in Alibaba's history

"Second place only behind

Fable 5" — this is how Alibaba described Qwen3.8-Max in the official Qwen blog in July 2026, one month before the public release.

Why the Open Release Intensifies the Technology Race

The release of Qwen3.8-Max with open weights is a strategic move amid the escalating technological competition between the US and China. Open weights mean that any developer in any country can study the model's architecture, adapt it, and deploy it on their own infrastructure. This significantly complicates attempts to limit the spread of Chinese AI technologies through export restrictions and sanctions.

The precedent is not new: when DeepSeek released open models comparable in capabilities to GPT-4o in early 2025, it sparked a wave of discussions in Washington about the real effectiveness of chip supply restrictions. Qwen3.8-Max continues the same trend: a powerful, open, globally competitive model from China. According to The Verge, each such release noticeably raises tensions both in Silicon Valley and among American policymakers.

Notably, Alibaba chose open access over a proprietary API — thus competing with Anthropic and OpenAI not only in quality but also in terms of accessibility.

What This Means

Alibaba is consistently reinforcing China's status among the leading players in the global AI race. The open release of Qwen3.8-Max increases pressure on American laboratories — OpenAI, Anthropic, Google DeepMind — both in terms of performance and openness of development. For developers around the world, the selection of powerful open-weight models continues to expand, while dependence on closed API services continues to shrink.

Frequently Asked Questions

What is Qwen3.8-Max?

Qwen3.8-Max is Alibaba's largest language model, released on August 11, 2026 with open weights. According to the company, it is comparable in performance to Anthropic's Fable 5 and OpenAI's flagships, and also surpasses Chinese competitor Kimi K3 from Moonshot AI.

Is Qwen3.8-Max available for free?

The model was released in open-weight format — the weights are open for downloading, fine-tuning, and use. Detailed commercial licensing terms are published on the official Qwen blog.