Alibaba unveils Qwen4-Max, claims it beats rival models on coding
Alibaba has unveiled Qwen4-Max, its new flagship AI model, claiming it outperforms rivals on coding benchmarks while offering aggressive pricing.
The new model runs on a sparse mixture-of-experts architecture with a 2.4-trillion-parameter backbone, designed to be both powerful and efficient. Unlike traditional models that activate all parameters for every task, Qwen4-Max selectively engages only relevant 'experts' for each request, reducing computational costs while maintaining high performance.
Technical capabilities
Qwen4-Max is multimodal, handling text, images, and video with a context window of up to 1 million tokens—roughly equivalent to the model's short-term memory. This allows it to process long documents or extensive codebases in a single session. Developers can access it through an API on QwenCloud, with compatibility for OpenAI and Anthropic protocols to ease integration.
Benchmark claims vs reality
Alibaba claims Qwen4-Max scored 75.7 on the DeepSWE v1.1 coding benchmark, outperforming Claude Opus 5.5 and competing with Kimi K3.1 Pro. However, independent rankings tell a different story, placing it behind competitors like Kimi while highlighting its favorable cost per task. The model uses approximately 50% fewer tokens than its predecessor, Qwen3.8-Max, with average task costs falling by over 60%.
Tokens are the chunks of text a model reads and writes, and they are also how most AI providers bill customers. Fewer tokens per job means a smaller invoice for the same work.
Pricing strategy
Alibaba has positioned Qwen4-Max as a cost-effective flagship model, pricing it at $1.50 per million input tokens and $4.50 per million output tokens. This pricing structure results in a cost per task of roughly $3.27, lower than many competitors. A Token Plan preview opened on November 11, 2026, and runs until November 25, though full access did not begin on launch day.
Market positioning
Rather than competing purely on capability, Alibaba is emphasizing efficiency and practical use, targeting professional productivity and coding tasks in sectors like healthcare, finance, and law. The company plans to release smaller open-weight models in the Qwen4 series while keeping the flagship closed to protect its most valuable asset. For businesses, the combination of lower token usage and competitive pricing may be more significant than raw performance numbers, though the gap between Alibaba's claims and independent rankings warrants careful evaluation.