Qwen3.8 Max 0902 keeps 1M context at $2/M

Alibaba’s Qwen3.8-Max-0902 snapshot keeps a 1-million-token context window and lists $2 per million input tokens and $6 per million output tokens on QwenCloud. The QwenCloud model page exposes a 131K maximum output and tool calling, while Alibaba’s Model Studio pricing table lists the same model at CNY 12/CNY 36 in its global regions; pricing varies by endpoint. Builders comparing Qwen’s open-model ecosystem and deployment ladder should treat this as a new snapshot to benchmark for long-context agent workloads, not a license to assume the 1M window is economical at full utilization.