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
Qwen3.8 Explainer
#mtp-qwen38-explainer *{box-sizing:border-box!important;margin:0;padding:0}
#mtp-qwen38-explainer{
background:#0b0b0b!important;color:#e6e6e6!important;
font-family:-apple-system,BlinkMacSystemFont,”Segoe UI”,Roboto,Helvetica,Arial,sans-serif!important;
border:1px solid #1e1e1e!important;border-radius:14px!important;overflow:hidden!important;
max-width:860px;margin:0 auto!important;line-height:1.5;
}
#mtp-qwen38-explainer .hd{padding:22px 24px 18px;border-bottom:1px solid #1a1a1a;background:linear-gradient(180deg,#111 0%,#0b0b0b 100%)}
#mtp-qwen38-explainer .kick{color:#76B900;font-size:11px;letter-spacing:2px;text-transform:uppercase;font-weight:700;margin-bottom:8px}
#mtp-qwen38-explainer h2{font-size:21px;color:#fff;font-weight:700;line-height:1.25}
#mtp-qwen38-explainer .sub{color:#8a8a8a;font-size:13px;margin-top:7px}
#mtp-qwen38-explainer .body{padding:20px 24px 8px}
#mtp-qwen38-explainer .tabs{display:flex;gap:8px;margin-bottom:18px;flex-wrap:wrap}
#mtp-qwen38-explainer .tab{flex:1;min-width:120px;cursor:pointer;padding:11px 10px;border-radius:9px;border:1px solid #242424;background:#111;color:#bdbdbd;font-size:13px;font-weight:600;text-align:center;transition:.15s}
#mtp-qwen38-explainer .tab:hover{border-color:#3a3a3a}
#mtp-qwen38-explainer .tab.on{background:rgba(118,185,0,.12);border-color:#76B900;color:#a6e000}
#mtp-qwen38-explainer .panel{display:none}
#mtp-qwen38-explainer .panel.on{display:block;animation:mtpf .25s ease}
@keyframes mtpf{from{opacity:0;transform:translateY(4px)}to{opacity:1;transform:none}}
#mtp-qwen38-explainer .row{display:flex;gap:10px;padding:11px 13px;border-radius:9px;background:#111;border:1px solid #1c1c1c;margin-bottom:9px;align-items:flex-start}
#mtp-qwen38-explainer .row .ic{flex:0 0 auto;width:20px;height:20px;border-radius:50%;font-size:12px;font-weight:800;display:flex;align-items:center;justify-content:center;margin-top:1px}
#mtp-qwen38-explainer .ic.y{background:rgba(118,185,0,.15);color:#76B900}
#mtp-qwen38-explainer .ic.n{background:rgba(255,120,90,.13);color:#ff7a5a}
#mtp-qwen38-explainer .row .t{font-size:13.5px;color:#dcdcdc}
#mtp-qwen38-explainer .row .t b{color:#fff}
#mtp-qwen38-explainer .note{font-size:12px;color:#7c7c7c;margin:4px 2px 16px}
#mtp-qwen38-explainer .chart{margin:6px 0 6px}
#mtp-qwen38-explainer .bar{display:flex;align-items:center;gap:10px;margin-bottom:10px}
#mtp-qwen38-explainer .bar .lab{flex:0 0 128px;font-size:12.5px;color:#cfcfcf}
#mtp-qwen38-explainer .bar .lab span{display:block;font-size:10.5px;color:#777}
#mtp-qwen38-explainer .bar .track{flex:1;background:#151515;border-radius:6px;height:26px;position:relative;overflow:hidden}
#mtp-qwen38-explainer .bar .fill{height:100%;border-radius:6px;display:flex;align-items:center;justify-content:flex-end;padding-right:8px;font-size:11.5px;font-weight:700;color:#0b0b0b;transition:width .8s cubic-bezier(.2,.7,.2,1)}
#mtp-qwen38-explainer .fill.g{background:linear-gradient(90deg,#4a7500,#76B900)}
#mtp-qwen38-explainer .fill.d{background:linear-gradient(90deg,#2c2c2c,#4d4d4d);color:#e6e6e6}
#mtp-qwen38-explainer .fill.q{background:linear-gradient(90deg,#5c8f00,#9be000);box-shadow:0 0 14px rgba(118,185,0,.4)}
#mtp-qwen38-explainer .calc{background:#0f0f0f;border:1px solid #1e1e1e;border-radius:11px;padding:16px}
#mtp-qwen38-explainer .seg{display:flex;gap:6px;margin:10px 0 14px}
#mtp-qwen38-explainer .seg button{flex:1;cursor:pointer;padding:9px 6px;border-radius:8px;border:1px solid #262626;background:#141414;color:#bcbcbc;font-size:12.5px;font-weight:600;transition:.15s}
#mtp-qwen38-explainer .seg button.on{background:rgba(118,185,0,.14);border-color:#76B900;color:#a6e000}
#mtp-qwen38-explainer .out{display:flex;gap:12px;flex-wrap:wrap}
#mtp-qwen38-explainer .stat{flex:1;min-width:150px;background:#111;border:1px solid #1c1c1c;border-radius:9px;padding:13px 15px}
#mtp-qwen38-explainer .stat .n{font-size:24px;font-weight:800;color:#76B900;line-height:1.1}
#mtp-qwen38-explainer .stat .k{font-size:11px;color:#8a8a8a;text-transform:uppercase;letter-spacing:.5px;margin-top:5px}
#mtp-qwen38-explainer .warn{margin-top:12px;font-size:12px;color:#e0b050;background:rgba(224,176,80,.08);border:1px solid rgba(224,176,80,.2);border-radius:8px;padding:9px 11px}
#mtp-qwen38-explainer .bench{display:grid;grid-template-columns:repeat(3,1fr);gap:10px}
#mtp-qwen38-explainer .bcard{background:#111;border:1px solid #1c1c1c;border-radius:9px;padding:13px}
#mtp-qwen38-explainer .bcard .bn{font-size:22px;font-weight:800;color:#fff}
#mtp-qwen38-explainer .bcard .bl{font-size:11px;color:#8a8a8a;margin-top:3px}
#mtp-qwen38-explainer .ft{padding:13px 24px;border-top:1px solid #1a1a1a;display:flex;justify-content:space-between;align-items:center;flex-wrap:wrap;gap:8px}
#mtp-qwen38-explainer .ft .mk{font-size:11.5px;color:#76B900;font-weight:700}
#mtp-qwen38-explainer .ft .dt{font-size:11px;color:#666}
@media(max-width:640px){
#mtp-qwen38-explainer .bench{grid-template-columns:1fr}
#mtp-qwen38-explainer .bar .lab{flex:0 0 96px}
#mtp-qwen38-explainer h2{font-size:18px}
}
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?
FP16 (16-bit)
FP8 (8-bit)
INT4 (4-bit)
—Weights only
—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.
92.4GPQA Diamond
80.4%SWE-bench Verified
69.7Terminal-Bench 2.0
1MContext window (tokens)
$1.25Per 1M input tokens
$3.75Per 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
(function(){
var root=document.getElementById(‘mtp-qwen38-explainer’);
var tabs=root.querySelectorAll(‘.tab’),panels=root.querySelectorAll(‘.panel’);
tabs.forEach(function(t){t.addEventListener(‘click’,function(){
tabs.forEach(function(x){x.classList.remove(‘on’)});t.classList.add(‘on’);
panels.forEach(function(p){p.classList.remove(‘on’)});
root.querySelector(‘#’+t.dataset.p).classList.add(‘on’);
if(t.dataset.p===’p2′)drawChart();
postSize();
})});
var models=[
{n:’Kimi K3′,s:’Moonshot AI’,v:2.8,cls:’g’},
{n:’Qwen3.8′,s:’Alibaba (claimed)’,v:2.4,cls:’q’},
{n:’DeepSeek V4′,s:’DeepSeek’,v:1.6,cls:’g’},
{n:’Inkling’,s:’Thinking Machines’,v:0.975,cls:’g’},
{n:’Qwen3.7-Max’,s:’Alibaba (est.)’,v:1.0,cls:’d’}
];
var drawn=false;
function drawChart(){
if(drawn)return;drawn=true;
var max=2.8,html=”;
models.forEach(function(m){
var pct=(m.v/max*100).toFixed(1);
html+=”+m.n+”+m.s+’