Alibaba releases Qwen3.8-Max: 2.4T-parameter AI model challenges Moonshot

Alibaba has officially launched its largest and most capable artificial intelligence model to date labelled as "Qwen3.8-Max."

As reported by Reuters, the Chinese tech giant unveiled its largest model on Monday to challenge newly launched Moonshot AI.

The launch comes as Chinese tech companies-a huge force in open-weight AI models are globally locked in a fierce and fast-moving battle to build more powerful systems without making them prohibitively expensive to run.

Key details about Qwen3.8-Max.:

The model boasts 2.4 trillion parameters, placing it just behind domestic rival Moonshot AI’s Kimi K3 model which features 2.8 trillion parameters.

To keep operational costs and response latency manageable, the model utilizes an MoE design that routes tasks through specialized subsystems, activating only 95 billion parameters at any given time.

Multimodal & Context Window:

Qwen3.8-Max handles text, images, and video, supporting a massive context window of up to 1 million tokens at once—allowing it to ingest extensive codebases, legal documents, or hundreds of pages of text in a single go.

By contrast, OpenAI, Anthropic and Google do not publish parameter count for their closed-source models.

Qwen3.8-Max was unveiled on crowdsourced, model-comparison platform ‌Arena.AI, where it immediately became the highest-ranking Chinese model in terms of text models, though it still lags Claude Fable 5 and three Opus variants which are all from Anthropic.

But on Arena. Ai’s leaderboard for AI models that analyse images and other visual material, Qwen3.8-Max ranked second globally, only behind a Claude Fable 5 variant.

Performance & Rankings:

Qwen3.8-Max was unveiled on crowdsourced. Upon its debut on the crowdsourced evaluation platform Arena.AI, it quickly secured the highest ranking among Chinese text models and claimed the global second-place spot for multimodal vision-language models, trailing only variants of Anthropic's Claude.

Availability: 

The model is slated for formal commercial availability via Alibaba Cloud’s Model Studio platform.

However, a preview version of Qwen3.8-Max is currently available for early testing through Alibaba's Token Plan ecosystem and supported developer platforms, with a wider general release and open-source model weights rolling out soon.