LEGACY QWEN32K CONTEXTBATCH -50%REPLACED BY QWEN3 MAX
Qwen Max API Pricing
Qwen Max is the legacy proprietary Qwen 2.5 Max row kept for backward-compatible workloads. Alibaba still lists it at $1.60/M input and $6.40/M output, while Qwen3 Max is the current replacement. Checked by hand against alibabacloud.com.
Input - per 1M tokens
$1.60/M
Source Alibaba Cloud flat
Output - per 1M tokens
$6.40/M
Context 32K flat
Cache N/A
$1.60/M
Cache no dollar row not listed
Effective - agentic blend
$1.98/M
92/8 split - no cache
§ 01 / TERMINAL
Qwen Max API cost calculator.
Live calculator pre-loaded with current Qwen Max rates. Tweak spend, output mix, or cache assumptions and share the URL to share the calculation.
$ /mo
Workload split
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
§ 02 / SCENARIOS
Qwen Max cost on real workloads.
LEGACY
Invoice replay
$0.012/turn
CHATBOT
Older app assistant
$0.00992/turn
RAG
Compatibility query
$0.031/query
MIGRATION
A/B comparison
$0.017/test
Calibrated · measured on the vendor's tokenizer · 2026-06-10 Auto-counts as you type
Counts use a chars-per-token calibration measured on the vendor's own published tokenizer (Qwen/Qwen3.5-397B-A17B, 2026-06-10). English prose is typically within a few percent; code and non-Latin scripts tokenize heavier. For billing-exact counts use the vendor's count-tokens API.
Characters 385
Words 61
Tokens (estimated) 74 tokens
Cost as input · uncached $0.00012 USD
Cost as output · uncached $0.00047 USD
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| Qwen Max (2.5) Current | $1.60 | $6.40 | $1.98 agentic 92/8 | 32K | Legacy Qwen 2.5 compatibility |
| Qwen3 Max | $1.20 | $6.00 | $1.58 retired | 252K | Frontier Qwen proprietary reasoning |
| Qwen 3.5 Plus | $0.40 | $2.40 | $0.56 −3.5× | 1M | General Qwen production workloads |
| Qwen 3.5 Flash | $0.10 | $0.40 | $0.124 −16× | 1M | Cheap long-context Qwen traffic |
| Qwen3 235B A22B | $0.70 | $2.80 | $0.868 retired | 131K | Open MoE reasoning baseline |
| Qwen3 32B | $0.16 | $0.64 | $0.198 retired | 131K | Open 32B chat and reasoning |
| Qwen3 14B | $0.35 | $1.40 | $0.434 retired | 131K | Compact open Qwen reasoning |
| Qwen 3.5 397B A17B | $0.60 | $3.60 | $0.84 −2.4× | 256K | Open MoE frontier workloads |
| Qwen 3.5 122B A10B | $0.40 | $3.20 | $0.624 −3.2× | 256K | Open MoE production chat |
Qwen Max API pricing FAQ.
Practical pricing questions, separated from calculator assumptions and regional tiers.
Q · 01 What is Qwen Max priced at? +
Qwen Max is shown at
$1.6/M input and $6.4/M output in Alibaba Cloud Model Studio's International/Singapore deployment section.Q · 02 Does this page use International or Global pricing? +
This page uses Alibaba Cloud Model Studio International deployment pricing, where endpoint and data storage are in Singapore and inference resources are dynamically scheduled globally excluding Chinese Mainland. Global and Chinese Mainland sections can list different prices.
Q · 03 Is prompt caching priced separately? +
Alibaba marks context-cache support on some Qwen families, but this row does not publish a concrete cache-read dollar price. The calculator therefore treats cached input as the same
$1.6/M baseline instead of inventing a discount.Q · 04 How is the effective price calculated? +
AI//COST uses the same 92/8 agentic blend everywhere. With no separate cache-read price published for this row, Qwen Max's effective blended cost is
$1.98/M.Q · 05 Is there a free quota or batch discount? +
Alibaba lists a 90-day activation free quota for many International Qwen rows, but quota eligibility is model-specific. Batch Invocation is 50% off where the vendor row explicitly marks Batch support; this page only publishes the real-time list prices in the quote tiles.
Q · 06 Are regional prices different? +
Yes. Alibaba Cloud publishes separate International, Global, China (Hong Kong), EU, US, and Chinese Mainland deployment sections. AI//COST uses the International/Singapore baseline for Alibaba queue pages unless a page explicitly says otherwise.