Skip to main content
Alibaba

Qwen 3.6 Plus

Qwen 3.6 Plus is a large language model from Alibaba in the Qwen 3.6 family, offered with a 1,000,000-token context window and accessible through API providers.

Context 1.0M
Input from
$0.325 / 1M tokens
across 5 providers

API Pricing

Cheapest on OpenRouter — 27% below avg
ProviderInput / 1MOutput / 1MCached / 1M
$0.325$1.95-
$0.400$2.40$0.040
$0.500$3.00-
$0.500$3.00-
$0.500$3.00$0.050

Prices updated daily. Last check: Sep 25, 2026

Qwen 3.6 Plus pricing by provider

Input, output, and batch rates, plus alternatives, for one provider at a time.

Performance & Benchmarks

Source: Artificial Analysis →
Intelligence
27.0 / 100
Coding
54.5 / 100

Reasoning & Knowledge

  • GPQA Diamond88.2%
  • Humanity's Last Exam27.8%

Agentic & Tool Use

  • Terminal-Bench Hard43.9%
  • Terminal-Bench v2.161.4%
  • τ²-bench97.7%
  • τ-bench Banking20.8%

Instruction & Long Context

  • IFBench75.2%
  • Long-Context Reasoning78.3%

Benchmarks measured Sep 2026. Scores are independent evaluations, not vendor-reported.

Model Details

General

Creator
Alibaba
Family
Qwen 3.6
Context Window
1.0M
Modalities
Text

Capabilities

Tool Calling
No
Open Source
No
Aliases
qwen3.6-plus, qwen3.6-plus-preview, Qwen3.6-Plus, Qwen/Qwen3.6-Plus

Strengths & Limitations

Strengths

  • 1,000,000-token context window, allowing whole repositories or large document sets in a single request
  • Part of Alibaba's Qwen 3.6 generation, so it shares tokenizer and API conventions with sibling models in that family
  • "Plus" tier positioning offers a middle option between the lighter and heaviest variants of the same generation
  • Exposed under multiple standard identifiers (qwen3.6-plus, Qwen/Qwen3.6-Plus), making it straightforward to locate across provider catalogs
  • Long-context capacity reduces the need for retrieval pipelines and chunking logic in document-heavy applications
  • Available through third-party inference providers, so per-token rates can be compared side by side in the pricing table on this page

Limitations

  • We do not currently track confirmed benchmark scores for Qwen 3.6 Plus, so capability comparisons require your own evaluation
  • Throughput and time-to-first-token measurements in our database are unpopulated for this model
  • Some provider listings use a -preview identifier, which may indicate behavior or availability changes between versions
  • Filling the full 1,000,000-token context is expensive in practice and can increase latency substantially regardless of provider
  • Modality support, tool calling and other API features are not fields we have confirmed for this model — check your provider's documentation

Key Features

•1,000,000-token context window
•Member of Alibaba's Qwen 3.6 model family
•"Plus" tier positioning within the Qwen 3.6 generation
•Multiple provider aliases: qwen3.6-plus, qwen3.6-plus-preview, Qwen3.6-Plus, Qwen/Qwen3.6-Plus
•Preview-tagged variant available under qwen3.6-plus-preview
•Accessible via hosted inference APIs with per-token billing
•Long-context ingestion for repository-scale code and multi-document analysis

About Qwen 3.6 Plus

Qwen 3.6 Plus is a model released by Alibaba as part of the Qwen 3.6 family. In Alibaba's Qwen naming convention, "Plus" designates a mid-to-upper tier positioned between the lighter, throughput-oriented variants and the largest models in the same generation, so it is generally aimed at workloads that need broad capability without defaulting to the heaviest option in the family. The most concretely documented specification we track for Qwen 3.6 Plus is its context window of 1,000,000 tokens. A context of that size allows entire code repositories, long document collections, extended chat histories, or large transcript sets to be placed directly in a single request rather than being chunked and retrieved piecewise. The model is exposed under several identifiers depending on the provider, including qwen3.6-plus, qwen3.6-plus-preview, Qwen3.6-Plus and Qwen/Qwen3.6-Plus, which is worth noting when wiring up an API client or comparing listings across vendors. Beyond the context window and family position, our database does not yet carry confirmed benchmark scores or throughput and latency measurements for this model — the performance figures we hold from Artificial Analysis are currently unpopulated. Readers evaluating Qwen 3.6 Plus against peers should therefore rely on their own task-level testing, and use the pricing table on this page to compare what individual providers charge for the same model identifier.

Common Use Cases

Qwen 3.6 Plus is best suited to work where input length is the binding constraint: reviewing or refactoring across a large codebase, question answering over long contracts, filings or technical manuals, summarizing lengthy meeting and support transcripts, and agent or assistant sessions that accumulate very long histories. The 1,000,000-token window means much of this can be handled without building a retrieval layer, which simplifies architecture for teams whose documents are large but whose query volume is moderate. As a "Plus"-tier model it is positioned as a general-purpose choice rather than a minimal-cost, high-volume classifier — for very large batch workloads, lighter variants in the same generation are usually the better fit, while teams needing maximum capability may prefer the larger models in Alibaba's lineup. Because we do not yet hold verified benchmark or latency data for this model, prototype on a representative sample of your own tasks before committing.

Frequently Asked Questions

How much does Qwen 3.6 Plus cost to use?

Pricing depends on which provider hosts the model and on the pricing type — input tokens, output tokens, cached input and long-context surcharges are often billed at different rates. Rates also change over time. See the pricing table on this page for current per-provider figures rather than relying on any fixed number.

What is Qwen 3.6 Plus best used for?

It fits long-context workloads: repository-scale code analysis, question answering over large document collections, long transcript summarization, and assistant sessions with extended histories. Its 1,000,000-token context window is the main reason to choose it over shorter-context alternatives.

How large is the context window?

Qwen 3.6 Plus supports a context window of 1,000,000 tokens. Note that individual providers sometimes cap the usable context below a model's maximum, so verify the limit in your provider's documentation.

How does Qwen 3.6 Plus compare to other models in the Qwen 3.6 family?

The "Plus" label places it in the middle-to-upper part of Alibaba's tier structure for the Qwen 3.6 generation — above the lighter, throughput-focused variants and below the largest option. If your workload is high-volume and simple, a lighter sibling is typically cheaper; if it demands maximum capability, the larger family member may be preferable.

Are there published benchmark results for Qwen 3.6 Plus?

Our database does not currently carry confirmed benchmark scores or throughput and latency measurements for this model. We would rather report nothing than report unverified numbers, so we recommend running your own evaluation on representative tasks.

What model identifier should I use in API calls?

It depends on the provider. Known identifiers include qwen3.6-plus, qwen3.6-plus-preview, Qwen3.6-Plus and Qwen/Qwen3.6-Plus. Check the provider entry in the pricing table and their API docs for the exact string.