Last verified
QWEN3.6 35B A3B256K CONTEXTTEXT + VISIONINTERNATIONAL PRICECNY CONVERTED

Qwen3.6 35B A3B API Pricing

Qwen3.6 35B A3B is an Alibaba Qwen3.6 model listed in the official International/Singapore pricing table. The verified row is CNY 2.810325/M input and CNY 16.86195/M output, shown here as $0.4142/M and $2.48/M using the 2026-06-30 CNY/USD reference rate.

Input - per 1M tokens
$0.41/M
Source Alibaba Model Studio International
Output - per 1M tokens
$2.48/M
Base row 0-256K International
Cached input - not itemized
$0.41/M
Cache discount exists, exact row not exposed not listed
Effective - agentic blend
$0.58/M
92/8 split - no cache discount assumed
§ 01 / TERMINAL

Run the numbers.

Live calculator pre-loaded with Qwen3.6 35B A3B's verified Alibaba International base row. Alibaba marks context-cache discount support, but this page does not assume a cache-read token price because the exact International cache token row is not itemized in the public pricing table.

$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
Open full calculator (all models · share URL · CSV) →
§ 02 / SCENARIOS

Real-world presets.

CODING

Agent implementation

$0.027/task
28,000 in - 6,000 out~3,772 units/$100
RAG

Long-doc answer

$0.050/answer
90,000 in - 5,000 out~2,012 units/$100
CHATBOT

Product assistant

$0.005/turn
5,000 in - 1,000 out~21,949 units/$100
BATCH

Knowledge summary

$0.018/doc
32,000 in - 1,800 out~5,641 units/$100
§ 03 / TOKENIZER

Paste text. See tokens. See cost.

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 390
Words 58
Tokens (estimated) 75 tokens
Cost as input · uncached $0.000031 USD
Cost as output · uncached $0.000186 USD
Cost as cached input $0.000031 USD
§ 04 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
Qwen3.6 35B A3B Current $0.41 $2.48 $0.58 agentic 92/8 256K MoE reasoning and coding
Qwen3.6 Flash $0.28 $1.66 $0.39 sibling 1M Fast multimodal agents
Qwen3.6 35B A3B Current $0.41 $2.48 $0.58 agentic 92/8 256K MoE reasoning and coding
Qwen3.6 27B $0.66 $3.98 $0.93 sibling 256K Dense multimodal reasoning

Frequently asked.

Practical Qwen3.6 35B A3B pricing questions, with Alibaba's CNY International row kept separate from USD conversion assumptions.

Q · 01 What is Qwen3.6 35B A3B priced at? +
Qwen3.6 35B A3B is listed at CNY 2.810325/M input and CNY 16.86195/M output in Alibaba Model Studio's International/Singapore table. Converted at 1 CNY = $0.14737, that is $0.4142/M input and $2.48/M output.
Q · 02 Why are the prices converted from CNY? +
Alibaba publishes the International Model Studio rows in CNY. AI//COST keeps the original CNY numbers in the page copy and snapshot notes, then converts to USD for cross-provider comparisons.
Q · 03 Is prompt caching priced separately? +
The pricing table marks context-cache discount support for this model family, but the public International row does not expose an exact cache-read token rate. Cached input is therefore kept equal to fresh input instead of inventing a discount.
Q · 04 Which API model ID should I use? +
Use qwen3.6-35b-a3b unless Alibaba's model catalogue tells you to pin a dated snapshot. Re-check the official Model Studio page before production deployment.
Q · 05 What is this model best for? +
It is a Qwen3.6 mixture-of-experts model positioned for reasoning, coding, STEM work, and native vision-language tasks. The page uses the base International row; very long prompts may use a higher long-context price if Alibaba routes them into a separate published band.
Y
Reviewed by Yaroslav Vikhariev Founder - AI//COST - Pricing pulled daily from Alibaba Model Studio - Last verified Jun 30, 2026
Methodology Report a correction More by Y.V.