Qwen3.8 Max Preview
Alibaba previewed Qwen3.8 Max on 19 Jul 2026 at the World AI Conference in Shanghai: a 2.4-trillion-parameter sparse MoE with text, image, video, and document input and a 1M-token context, which Alibaba claims is second only to Claude Fable 5 among frontier models. Treat that claim with caution — it shipped with no benchmark table, no model card, no license, no disclosed active-parameter count, and every figure so far comes from Alibaba's internal evals rather than Artificial Analysis or LMArena. Access is through Alibaba's credit-based Token Plan (Lite $6 / 2,500 credits per 7 days up to Pro $68 / 40,000 credits) at 10% of standard rates during preview; no dedicated per-token API rate has been published, so the prices shown here mirror Qwen3.7 Max and are a placeholder, not a quote.
1M
tokens
33K
tokens
$2.5
per 1M tokens
$7.5
per 1M tokens
86
tokens/sec
Jul 2026
2026-07-19
$5.00
per 1M tokens
19.0
quality per $
Capabilities
Benchmarks
About Qwen3.8 Max Preview
Qwen3.8 Max Preview is a flagship AI model by Alibaba Cloud, released on July 19, 2026. It supports a context window of 1000K tokens and can generate up to 33K output tokens.
At $2.5 per million input tokens and $7.5 per million output tokens, its blended cost of $5.00/1M tokens places it in the premium pricing tier. Its value score of 19.0 reflects the balance of quality and cost.
Using Qwen3.8 Max Preview with Swfte
Access Qwen3.8 Max Preview through Swfte Connect, our unified LLM gateway. Connect gives you a single API for 50+ models, with automatic routing, cost optimization, and fallback handling. You can also try Qwen3.8 Max Preview in our AI Playground before integrating.