Alibaba’s Qwen3.8-Max, a 2.4-trillion-parameter mixture-of-experts model with 95 billion active parameters and 1M-token context, launched via API in early August 2026 after a July preview and remains the current flagship on benchmarks. Traders see the next 3.9+ iteration as imminent because Alibaba has maintained a rapid release cadence for the Qwen series, with incremental Max refreshes and supporting open-weight variants appearing every few months. Competitive pressure from other Chinese labs and the shift toward Qwen 4 training further accelerate expectations for an update focused on coding, agentic tasks, and multimodal capabilities. The Max tier has stayed largely API-only or under restrictive licensing, while smaller models receive broader open releases, shaping sentiment around timing and access. No official 3.9 announcement has occurred as of mid-November 2026, but markets price strong odds of a near-term drop.