See how ChatGPT, Perplexity and Google AI Overviews describe you today
Free AI Visibility AuditQwen3.6 Max Preview is Alibaba Cloud's proprietary frontier text model, a preview release of the top tier of the Qwen line built for agentic coding and tool use.
QwenVerified 25/09/2026Released 27/04/202601 / decision
Each figure sits next to the middle of the field, so it can be read as dear or cheap, wide or narrow, rather than floating on its own. Compared against the 231 models we track, not against an absolute standard.
Capability
Not measured
No published benchmark scores yet
Input / 1M tokens
$1.03
field median $0.42
Output / 1M tokens
$6.16
field median $1.81
Context
262K
field median 500K
02 / overview
Four questions, answered separately, because somebody arrives at one of them rather than at the top of the page.
The job it was built for, and the job it is not for.
Qwen3.6 Max Preview is a closed-weight frontier model aimed at agentic work: writing and fixing code, calling tools, and running multi-step tasks rather than answering one question at a time. It is a text model, so there is no image, audio or video input to work with. As a preview build it is a place to test what the next Max generation does, not a fixed target to standardise a production stack on.
Where it sits in its line, and when a sibling is the better pick.
Qwen3.6 Max Preview first appeared in late April 2026 as the preview of the Max tier, the largest and most capable rung of the Qwen family, above the open-weight models Qwen ships alongside it. Reach for it when a job needs long-running agentic coding or tool orchestration and the smaller Qwen models keep dropping the thread. It is the wrong pick for high-volume classification, short chat replies, or anything needing images or audio, where a smaller or multimodal sibling costs less and does the job.
The API, the apps it powers, and what its limits let you do.
Qwen3.6 Max Preview is reached through the Qwen API on Alibaba Cloud and through the Qwen chat app, and it surfaces in the aggregators and coding tools that carry Qwen models. The context window holds a large working set, enough for a substantial repository slice plus the transcript of a long tool-calling session, and the generous output ceiling means it can return a full file or a long patch in one response rather than being paged out. Everything in and out is text.
What changes because this exists, or why it does not.
Agentic runs burn context fast, every tool call and every returned result lands back in the prompt, and the combination of a wide window and a high output limit here means fewer sessions that die halfway through a refactor. For teams already inside the Alibaba Cloud ecosystem, it removes the reason to route the hardest coding tasks out to a Western frontier model.
Follows Qwen3.6 Flash. Superseded by Qwen3.7 Max. See the whole line.
03 / evidence
Split by category, so a strong number never hides a thin evidence base. Verified means we read it on the benchmark's own published results; a provider's claim about its own model is shown and labelled rather than dropped.
No published benchmark scores for this model yet.
We publish a score only where we can link the result to where it was published. Until a benchmark result for this model exists in a source we read, this section stays empty rather than being filled with a provider's marketing figure.
How we decide what counts as evidence04 / ledger
Every published row, grouped by category, each compared with the best published score on the same benchmark. 'Is 64% good' is a question nobody can answer; '26 points behind the leader' is one anybody can.
Nothing in the ledger yet.
Each row here carries a score, the benchmark it came from, what the leading model scored on the same test, and a link to the published result. Rows appear as results are published and read.
How we decide what counts as evidence05 / capability
Where this model is strong, and against how many peers. Ranks are against models with evidence in that category, not against everything we track: ranking against models nobody tested would rank who published, not who is better.
No category scores to shape yet.
A category score is the weighted mean of the benchmarks published for it. With no published rows there is nothing to average, and an empty chart drawn at zero would say something false.
How we decide what counts as evidence06 / cost
List API rates as last read from the provider, with the source on every row, plus every change we have recorded since we started tracking it.
| Charge | Price | Unit | Read on | Source |
|---|---|---|---|---|
| Input | $1.03 | per 1M tokens | 2026-09-24 | Check |
| Output | $6.16 | per 1M tokens | 2026-09-24 | Check |
Qwen3.6 Max Preview sits in the mid range for a frontier-class model, cheaper per call than the flagship tiers from the big US labs while asking noticeably more than Qwen's own open-weight options. Output costs several times what input does, as usual, so the long generations it is good at are where the bill actually accumulates, and it pays to keep agent loops from regenerating whole files they could patch.
| Workload | Input / month | Output / month | Cost |
|---|---|---|---|
| A small product team | 20M tokens | 5M tokens | $51.35 |
| A busy support assistant | 200M tokens | 40M tokens | $451.88 |
| A document pipeline | 1000M tokens | 100M tokens | $1,643.20 |
List API rates, no caching and no batch discount, which both providers offer and which change the answer a great deal. Treat these as the ceiling, not the bill.
07 / specs
As listed by the provider's own catalogue and re-read every few hours. Anything absent is absent there too.
| Context window | 262,144 tokens |
|---|---|
| Maximum output | 65,536 tokens |
| Modalities | text |
| Released | 27/04/2026 |
| Status | Current |
| Catalogue identifier | qwen/qwen3.6-max-preview |
08 / lineage
What this model replaced, what replaced it, and what else its provider has in the field.
Also from Qwen
09 / line
Every model in this family in release order, so a page from eight months ago says in one glance that two newer ones exist.
Came before
← Qwen3.6 FlashCame after
Qwen3.7 Max →Ordered by release date and worked out from the naming, so a new member slots in as soon as its page exists. A retired model keeps its page and its place in the line.
10 / notes
What this model changes for a brand trying to be cited in AI answers, and every change we have logged since it launched.
What it changes for you
If your buyers include engineering teams in China or across Asia, Qwen answers are a surface your brand either appears in or does not, and it is a different retrieval and training mix from the Western assistants most SaaS teams monitor. The practical move is to check how your category and your product are described when the question is asked through Qwen, rather than assuming what holds in one assistant holds everywhere. Documentation that is clear, structured and publicly crawlable travels furthest here, as it does everywhere else.
Buyers meet Qwen3.6 Max Preview mainly in the Qwen chat app and through the Alibaba Cloud API, and developers meet it inside coding agents and IDE extensions that let you pick a model. Its reach is strongest with teams operating in China and across Asia, and with anyone comparing frontier options on price rather than defaulting to the familiar names.
Nothing published here yet. Changes appear within hours of a provider announcing them.
11 / questions
The things people ask about this model, answered from what is on this page rather than from anywhere else.
Models change what gets cited. We measure whether AI answers name your brand or your competitors across every assistant, and show you what to change.