OpenRouter has added Alibaba's Qwen4 Max, which the Qwen team positions for coding, office work and long-horizon agents.
OpenRouter says Alibaba's Qwen4 Max is now available through its model-routing service. In the provider's announcement on X, it describes Qwen4 Max as the first model in the Qwen4 series and says it is intended for coding, office work and long-horizon agents. The announcement adds another route for customers who want to call the model through OpenRouter rather than configure a separate vendor integration.
Alibaba released Qwen4 Max more broadly on November 16. The company describes a new architecture with 2.4 trillion parameters, native vision and a one-million-token context window. On Alibaba's own Work360 evaluation, the reported score rose from 47.74 for Qwen3.8 Max to 52.69, while average cost per task fell from $4.43 to $1.67, a 62% reduction. These are vendor-reported benchmark results, not independent measurements.
For QwenCloud, Startup Fortune reports input pricing of $1.50 and output pricing of $4.50 per million tokens, each 25% below Qwen3.8 Max, with implicit caching at $0.075 per million tokens. Those QwenCloud figures should not be treated as OpenRouter's rates. OpenRouter's post does not set out pricing, rate limits or regional restrictions, so customers should check the live model entry before switching production workloads.