Alibaba Qwen has created Qwen3.8-Max The open weights are confirmed to be widely available, and will ship the following week. Qwen3.8-277B is a second checkpoint. going open-weights. Qwen3.8-Max is a 2.4-trillion-parameter mixture-of-experts model. The model accepts video, images and text as input.
Is it deployable
It depends which artifact (or surface) you’re applying.
Today, any size company can use the hosted API. This is a hosted API. OpenAI- and DashScope-compatibleThe integration of the model is done by changing both base-URLs and IDs. Open weights is a completely different issue. The checkpoint, with a total of 2.4T parameters is an artifact from a datacenter that has multiple nodes. Alibaba did not disclose the number of activated parameters. It is not possible to model the cost of service. The checkpoint for GPUs on premises is Qwen3.8.27B.
Four industries are represented by the published features. Software engineering, document reviews for legal and financial documents, media, e-commerce, and design are the four industries.
Code agents at repository scale and knowledge bases for long documents can be used. The same technology can be used for long video indexing, structured information extraction and research assistants that work in multiple steps.
Interactive explainer
“+row[0]+’‘+(vals[qi]!==null&&vals[fi]!==null?(vals[qi]>vals[fi]?’above’:vals[qi]
‘;
C.m.forEach(function(mn,j));
el.innerHTML=h;L.appendChild(el)});
q(‘#bmw’).textContent=”Qwen3.8-Max above Fable5 on “”+cmp+ comparable rows” is a win-win situation.
setTimeout(function(){qa(‘#bml .f’).forEach(function(f){f.style.width=f.dataset.w+’%’})},40);
setTimeout(rs,140);}
/* RL curve */
Var RL=[[0,.474],[500,.469],[1000,.487],[1500,.586],[2000,.606],[2500,.622],[3000,.616],[3500,.647],[4000,.725],[4500,.719],[5000,.689]];
var “rlDone”=”false”;
function paintRl(){
if(rlDone)return;rlDone=true;
var g=q(‘#rlg’),X=function(v){return 44+v/5000*570},Y=function(v){return 210-(v-.44)/(.76-.44)*180};
function mk(t,a){var e=d.createElementNS(‘http://www.w3.org/2000/svg’,t);
for(var k in a)e.setAttribute(k,a[k]);g.appendChild(e);return e}
[.45,.5,.55,.6,.65,.7,.75].forEach(function(v){
mk(‘line’,{x1:44,y1:Y(v),x2:614,y2:Y(v),stroke:’#1d1830′,’stroke-width’:1});
var t=mk(’text’,{x:6,y:Y(v)+3,class:’lbl’});t.textContent=v.toFixed(2)});
mk(‘line’,{x1:44,y1:Y(.474),x2:614,y2:Y(.474),stroke:’#4a4170′,’stroke-width’:1,’stroke-dasharray’:’4 4′});
var sb=mk(’text’,{x:470,y:Y(.474)+14,class:’lbl’});sb.textContent=”SFT baseline 0.474″;
var pts=RL.map(function(p){return X(p[0])+’,’+Y(p[1])}).join(‘ ‘);
var pl=mk(‘polyline’,{points:pts,fill:’none’,stroke:’#7C5CFF’,’stroke-width’:2.6,’stroke-linejoin’:’round’,’stroke-linecap’:’round’});
var len=pl.getTotalLength?pl.getTotalLength():2000;
pl.setAttribute(‘stroke-dasharray’,len);pl.setAttribute(‘stroke-dashoffset’,len);
var an=d.createElementNS(‘http://www.w3.org/2000/svg’,’animate’);
an.setAttribute(‘attributeName’,’stroke-dashoffset’);an.setAttribute(‘from’,len);an.setAttribute(’to’,0);
an.setAttribute(‘dur’,’1.5s’);an.setAttribute(‘fill’,’freeze’);pl.appendChild(an);
RL.forEach(function(p,i){
var pk=p[1]===.725;
var c=mk(‘circle’,{cvar text=mk(),x:X (p[0]),cy:Y(p[1]),r:pk?6:3.6,fill:pk?’#B49CFF’:’#7C5CFF’,opacity:0});
var a2=d.createElementNS(‘http://www.w3.org/2000/svg’,’animate’);
a2.setAttribute(‘attributeName’,’opacity’);a2.setAttribute(‘from’,0);a2.setAttribute(’to’,1);
a2.setAttribute(‘dur’,’.3s’);a2.setAttribute(‘begin’,(i*0.13)+’s’);a2.setAttribute(‘fill’,’freeze’);c.appendChild(a2);
var t=mk(’text’,{x:X(p[0]),y:Y(p[1])-12,’text-anchor’:’middle’,class:’lbl’,fill:pk?’#B49CFF’:’#8b81ab’,opacity:0});
t.textContent=p[1].toFixed(3);
var a3=d.createElementNS(‘http://www.w3.org/2000/svg’,’animate’);
a3.setAttribute(‘attributeName’,’opacity’);a3.setAttribute(‘from’,0);a3.setAttribute(’to’,1);
a3.setAttribute(‘dur’,’.3s’);a3.setAttribute(‘begin’,(i*0.13)+’s’);a3.setAttribute(‘fill’,’freeze’);t.appendChild(a3);
if(i%2===0){var xl=mk(’text’,{x:X(p[0]),y:228,’text-anchor’:’middle’,class:’lbl’});xl.textContent=p[0]}});
var pk=mk(’text’,{x:X(4000),y:Y(.725)-26,’text-anchor’:’middle’,class:’lbl-b’,fill:’#B49CFF’});
pk.textContent=”peak”;
setTimeout(rs,200)}
/* cross-harness */
Var HZ=[
{n:’CoWorkBench’,b:[[‘Fable5 (OpenClaw)’,75.9,0],[‘Opus4.8 (OpenClaw)’,72.3,0],[‘Qwen3.7-Max (OpenClaw)’,64.6,0],[‘QwenWork’,73.2,1],[‘Claude Code’,74.6,1],[‘Codex’,75.8,1],[‘OpenClaw’,74.8,1],[‘Hermes’,73.3,1]]},
{n:’WorkspaceBench’,b:[[‘Fable5 (OpenClaw)’,68.7,0],[‘Opus4.8 (OpenClaw)’,66.8,0],[‘Qwen3.7-Max (OpenClaw)’,61.4,0],[‘QwenWork’,67.0,1],[‘Claude Code’,67.3,1],[‘Codex’,67.6,1],[‘OpenClaw’,67.7,1],[‘Hermes’,71.2,1]]},
{n:’JobBench’,b:[[‘Fable5 (OpenCode)’,57.4,0],[‘Opus4.8 (OpenCode)’,48.4,0],[‘Qwen3.7-Max (OpenCode)’,31.3,0],[‘QwenWork’,58.0,1],[‘Claude Code’,59.0,1],[‘Codex’,57.5,1],[‘OpenClaw’,59.8,1],[‘Hermes’,59.6,1],[‘OpenCode’,53.4,1]]}];
HZ.forEach(function(h){
var c=d.createElement(‘div’);c.className=”hc”;
The var s=”
“+h.n+’
‘;
var mx=Math.max.apply(null,h.b.map(function(x){return x[1]}));
h.b.forEach(function(b){s+=’
‘+
‘‘+b[0]+’‘+
‘‘+b[1].toFixed(1)+’
‘});
c.innerHTML=s;q(‘#hzl’).appendChild(c)});
/* deploy */
Var DEP=[
‘Path: hosted API only. Call model ID qwen3.8-max through the DashScope or OpenAI-compatible endpoint. Self-hosting 2.4T weights is out of reach. Best fits: coding agents, document and video ingestion, research assistants. Owner: the founding engineer or a solo AI engineer.’,
‘Path: API first, 27B for anything sensitive. Use the hosted flagship for long-horizon agent work and keep Qwen3.8-27B on your own GPUs for PII-bearing or high-volume routing. Best fits: support automation, analytics copilots, internal tooling. Owner: an ML platform team.’,
‘Path: both, with an evaluation gate. Multi-node clusters can host the open weights once activated parameters and license are published. Until then run the API behind your own harness. The cross-harness numbers suggest portability across Claude Code, Codex and OpenClaw. Owner: an applied AI group reporting to the CTO.’,
‘Path: wait, then take 27B. No license file has been published for Qwen3.8-Max, so procurement cannot clear it yet. Air-gapped deployment realistically means Qwen3.8-27B. Best fits: healthcare, defense, public sector. Owner: security architecture plus a data science lead.’];
qa(‘#who button’).forEach(function(b){b.onclick=function(){
qa(‘#who button’).forEach(function(x){x.classList.remove(‘on’)});
b.classList.add(‘on’);q(‘#dep’).innerHTML=DEP[+b.dataset.w];setTimeout(rs,60)}});
q(‘#dep’).innerHTML=DEP[0];
/* evidence */
Var EV=[[
‘Full benchmark table across coding, agent, reasoning, multimodal, document, video’,
‘2.4 trillion total parameters, mixture-of-experts’,
‘1M context; 991K max input, 131K max output, 262K max reasoning’,
‘Text, image and video input; text output’,
‘$2.00 input, $6.00 output, $0.25 implicit cache read per 1M tokens’,
‘Built-in code_interpreter, web_search, web_extractor, t2i_search, i2i_search’,
‘RL scaling curve, including the decline past the 4,000-environment peak’,
‘Cross-harness results on QwenWork, Claude Code, Codex, OpenClaw, Hermes’,
‘Open weights for Qwen3.8-Max and Qwen3.8-27B announced for next week’
],[
‘Model card’,
‘License file’,
‘Activated parameters per token’,
‘Independent evaluation from Artificial Analysis or LMArena’,
‘Harness, attempt count and reasoning settings behind each score’,
‘Note: the multimodal table compares against Qwen3.7-Plus, not Qwen3.7-Max’
]];
var evMode=0
Function paint(){var L=q(‘#evl’);L.innerHTML=”;
The EV[evMode].forEach(function(t,i){var e=d.createElement(‘div’);e.className=”q-ev”;
e.style.animationDelay=(i*55)+’ms’;
e.innerHTML=’‘+(evMode?’?’:’✓’)+’‘+t+’‘;
L.appendChild(e)});setTimeout(rs,120)}
qa(‘#evp button’).forEach(function(b){b.onclick=function(){
qa(‘#evp button’).forEach(function(x){x.classList.remove(‘on’)});
b.classList.add(‘on’);evMode=+b.dataset.e;PaintEv()}});
paintEv();paintBm();
window.addEventListener(‘load’,rs);setTimeout(rs,300);setTimeout(rs,900);
})();
” style=”width:100%;height:600px;border:0;overflow:hidden;display:block” scrolling=”You can also find out more about the following:” loading=”lazy” title=”Qwen3.8-Max Interactive Explainer”>
What’s technically available
It is important to note that the word “you” means “you”. model page Lists a context window with 1M tokens. When thinking is active, the maximum input drops to 983K tokens. The maximum token output in either mode is 131K, with a budget of 262K for reasoning. There are rate limitations of 2M tokens and 15K request per minute.
Prices are $2.00 per 1000 input tokens and $ 6.00 per 1000 output tokens. Implicit cache Reads are $0.25 for 1M tokens. The explicit creation of cache costs $2.50, and the explicit readouts cost $0.17 for 1M tokens. Input that is cached costs eight times less than input from scratch. Cost is driven more by prefix stability than prompt length.
Included in the supported capabilities are function calling, structured outputs, batches, prefix completion Five built-in tools are available to fine-tune the Responses API: There are five tools that come with the Responses API. code_interpreter, web_search, web_extractor, t2i_search You can also find out more about the following: i2i_search.
Enjoy the Performance
Alibaba released a complete benchmark table along with the release. Qwen3.8 Max scored 86.6 points on Terminal Bench 2.1. It was ahead of Claude Opus 4.8 as well as Claude Fable 5. Fable 5 scored 80.0. It scores 67.7 in SWE-bench pro and 73.5 in FrontierSWE. PaperBench, at 93.0 points, is the leader. IFBench follows at 82.8. GPQA Diamond reaches 92.6, a marginal increase over Qwen3.7 Max’s previous 92.4. Not reasoning but multimodal and agentsic are where the gains are most evident. The most popular vision rows are OSWorld Verified at 86.1 and Parametric CAD Bench, 91.5. OmniDocBench, 1.5, is 92.1. It is a big leap from DeepSWE, which went from 21.6 points to 56.6, FrontierSWE, 40.7 points to 73.5. JobBench also went from 31.3 point to 53.4. In any fair reading, two cautions are necessary. This multimodal table compares Qwen3.7 to the benchmark.You can also read more aboutQwen3.7-Max is the one that flatters generational differences, and not Qwen3.7. Alibaba’s RL curve peaks near 4,000 learning environments at 0.725, before declining to 0.719 and 0.689.
The Key Takeaways
- The Qwen3.8 Max is now available in general. It has a 1M context and a 2.4T parameter.
- Prices are $2 input, 6 outputs and 0.25 cashed input per 1000 tokens.
- Next week, open weights of Qwen3.8 Max and Qwen3.8 27B will be available.
- The benchmark tables, the licenses, and activated parameter counts have not been released.
- On-premise deployment is realistically achieved by using the 27B Checkpoint and not the Flagship.
Click here to find out more Technical details, API You can also find out more about the following: Qwen Studio. Also, feel free to follow us on Twitter Don’t forget about our 150k+ML SubReddit Subscribe now our Newsletter. Wait! What? now you can join us on telegram as well.
Want to promote your GitHub repo, Hugging Face page, Product release or Webinar?? Connect with us
Asif Razzaq, CEO of Marktechpost Media Inc. is a visionary engineer and entrepreneur who is dedicated to leveraging the power of Artificial Intelligence (AI) for the social good. Marktechpost was his most recent venture. This platform, which specializes in covering machine learning and deep-learning news, is both technically solid and understandable to a broad audience. This platform has over 2,000,000 monthly views which shows its popularity.

