Qwen 3.8 Max Is Here: How It Actually Compares to Claude Fable 5 and Kimi K3

Chinese technology and e-commerce conglomerate Alibaba has unveiled a new flagship artificial intelligence model, Qwen3.8-Max, a 2.4 trillion-parameter system it says can compete with some of the world's most advanced AI systems, including Anthropic's Fable 5. 

The launch comes days after Beijing-based Moonshot AI released its Kimi K3 model for public download, and sent Alibaba shares up more than 6% in Hong Kong as investors bet the launch signals real enterprise traction.

On Arena.AI, the crowdsourced model-comparison leaderboard, Qwen3.8-Max has already debuted as the top-ranked Chinese text model and the No. 2 system globally for multimodal (vision-and-language) tasks, trailing only a Claude Fable 5 variant.

It's also priced aggressively: roughly $2 per million input tokens and $6 per million output tokens, well below what Fable 5 charges.

Alibaba Bans Claude Code Over Backdoor That Tracked Users

Alibaba has banned staff from using Claude Code, calling it “high-risk software” over an alleged backdoor that flagged Chinese users.

TechloyDamilare Odedina

How does Qwen3.8-Max compare to Anthropic's Fable 5? 

Alibaba says Qwen3.8-Max performs on par with, and in some cases exceeds, Anthropic's Fable 5 across several text, reasoning, coding, and multimodal benchmarks.