Qwen3.8-Max is generally available today, August 3, topping Alibaba's line with a 1M-token context window at $2/$6 per million tokens, as reported by OpenRouter and model trackers; pricing varies by region and provider. Earlier versions of this story said 256K and about $2.50/$7.50.

The release continues one of the most regular cadences in the industry: Qwen3-Max in September 2025, Qwen3.5-Max this February, Qwen3.7-Max in May, and now 3.8-Max in August, roughly quarterly, every time. No successor has a published date; on cycle alone, the next Max would land around November-December, and our Model Guide labels that a low-confidence estimate.

Where Qwen3.8-Max fits

In our Benchmarks, Qwen3.8-Max is the steady mid-cycle flagship: best for multilingual workloads and Alibaba Cloud stacks, priced well under the Western flagship bracket. It now has to hold that position through an autumn in which Anthropic, OpenAI, Google and xAI all ship new generations in September.

Adopt on merit today; the Max line refreshes roughly quarterly, so re-evaluate next quarter rather than waiting now.

The Qwen family carries a Proven badge in the Model Guide: late-cycle, fully rolled out, no dated successor, and no reason to hold a workload that fits it today.