On July 19, Alibaba’s Qwen team previewed Qwen3.8-Max-Preview, the next flagship in the Qwen family. The research team describes it as a 2.4 trillion-parameter model, ‘second only to Fable 5’ among the systems it benchmarked. The preview is live now. The benchmark table, model card, and license are not.
The July 19th 2026 announcement landed during the World AI Conference (WAIC) in Shanghai. It also arrived two days after Moonshot AI released Kimi K3, a 2.8 trillion-parameter open-weight model. The timing is the story as much as the model.
This article separates what Alibaba confirmed from what it only claimed. Every performance figure below carries that caveat.
What Qwen announced
The Qwen account posted that Qwen3.8 is launching and going open-weight soon. It called the model ‘one of the most powerful available today, comparable to leading frontier systems.
The preview build is real and purchasable. Access runs through Alibaba’s Token Plan subscription. The preview is offered at 10% of standard pricing.
Qwen developer Shuai Bai added technical detail. He described Qwen3.8 as the team’s first multimodal model above 1 trillion parameters. It processes text, images, video, and documents. Alibaba team states the model should beat Qwen3.7-Max on coding, full-stack development, data analysis, and office workflows.
Interactive Explainer
Marktechpost Interactive Explainer
Qwen3.8-Max-Preview: what Alibaba confirmed, what it only claimed
A 2.4T-parameter multimodal preview shipped before any benchmark, model card, or license. Explore the facts below.
Confirmed vs Claimed
Parameter scale
Serving-cost calculator
Verified 3.7-Max baseline
✓
Preview is live and purchasable. Qwen3.8-Max-Preview is sold through Alibaba Token Plan, Qoder, and QoderWork at 10% of standard pricing.
✓
Sparse Mixture-of-Experts, multimodal. Developer Shuai Bai says it is the team’s first multimodal model above 1T parameters, handling images, video, and documents.
✓
OpenAI and Anthropic protocol compatibility. Existing coding agents can point at Qwen3.8 without rebuilding their harnesses.
✕
2.4 trillion parameters. This is Alibaba’s own figure. No model card or specification confirms it.
✕
“Second only to Fable 5.” No benchmark table has been published. The ranking rests on internal evaluation.
✕
Open weights “soon.” No date, no license, no Hugging Face repository. Alibaba’s last two Max flagships shipped closed.
✕
Active parameters per token. Undisclosed — the single number that decides real serving cost for a sparse MoE model.
Status as of July 19, 2026. Toggle a fact type using the tabs above.
Total parameters, publicly disclosed frontier models, July 2026. Qwen3.8’s 2.4T is Alibaba’s claim, not a verified figure.
Total count is not usable compute. A sparse MoE model activates only a fraction of these parameters per token.
If the full 2.4T model shipped open-weight, what would it take to load?
—
Nvidia H200 (141GB) cards
Estimate for weights alone; add roughly 20–30% for the KV cache and runtime overhead. Active-parameter serving would be far cheaper, but Alibaba has not published that number.
The verified predecessor. Every published Qwen3.8 “capability” number is really a Qwen3.7-Max number until Alibaba releases a benchmark table.
1M
Context window (tokens)
$3.75
Per 1M output tokens
Qwen3.7-Max, May 2026, closed weights. Alibaba’s historic edge has been price-to-performance, not topping a leaderboard.
Marktechpost
Data verified July 19, 2026 · Figures marked claimed are Alibaba’s own
