Last verified
30B MOE262K CONTEXTTHINKING ONLYINTERNATIONAL PRICETEXT + CODE

Qwen3 30B A3B Thinking 2507 API Pricing

Qwen3 30B A3B Thinking 2507 is an Alibaba Qwen3 text model priced from the Singapore/International row of Model Studio. The verified rate is $0.2168/M input and $2.6/M output, converted from CNY at 1 CNY = $0.14768 Qwen lists this checkpoint as 30.5B total / 3.3B active with thinking-only behavior.

Input - per 1M tokens
$0.22/M
Source Alibaba Model Studio flat
Output - per 1M tokens
$2.60/M
Mode thinking only flat
Cache N/A
$0.22/M
Cache not listed not listed
Effective - agentic blend
$0.41/M
92/8 split - no cache
§ 01 / TERMINAL

Run the numbers.

Live calculator pre-loaded with the verified International Qwen3 30B A3B Thinking 2507 rates. Use it for invoice checks, agent traces, and scenario planning before large Model Studio runs.

$ /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.

REASONING

Math solution

$0.009/problem
12,000 in - 2,500 out~10,989 units/$100
CODING

Patch plan

$0.022/task
28,000 in - 6,000 out~4,608 units/$100
RAG

Evidence synthesis

$0.020/query
65,000 in - 2,200 out~5,051 units/$100
AGENT

Research branch

$0.033/task
90,000 in - 5,000 out~3,077 units/$100
§ 03 / TAPE

Price history.

Qwen3 30B A3B Thinking 2507 is listed at $0.2168/M input and $2.6/M output on Alibaba's International Model Studio row.

Input · $0.22/M
Output · $2.6/M
Cached · $0.22/M
JUL 29 Qwen3 30B A3B Thinking 2507 model card created on Hugging FaceJUN 09 Live International pricing verified at $0.2168/M input and $2.6/M output
§ 04 / 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 413
Words 61
Tokens (estimated) 79 tokens
Cost as input · uncached $0.000017 USD
Cost as output · uncached $0.000205 USD
Cost as cached input $0.000017 USD
§ 05 / SHELF

Up against the shelf.

All models →
Model Input /M Output /M Effective blended Context Best for
Qwen3 30B A3B Thinking 2507 Current $0.22 $2.60 $0.41 agentic 92/8 262K Compact MoE reasoning
Qwen3 Next 80B A3B Thinking $0.16 $1.30 $0.25 thinking 262K Qwen3 sibling
Qwen3 Next 80B A3B Instruct $0.16 $1.30 $0.25 sibling 262K Qwen3 sibling
Qwen3 235B A22B Thinking 2507 $0.25 $2.49 $0.43 thinking 262K Qwen3 sibling
Qwen3 235B A22B Instruct 2507 $0.25 $1.00 $0.31 sibling 262K Qwen3 sibling
Qwen3 30B A3B Instruct 2507 $0.22 $0.87 $0.27 sibling 262K Qwen3 sibling
Qwen3 30B A3B $0.22 $0.87 $0.27 sibling 262K Qwen3 sibling
Qwen3 8B $0.20 $0.76 $0.24 sibling 262K Qwen3 sibling
Qwen3.7 Plus $0.44 $1.77 $0.55 sibling 1M Current Plus tier
Qwen3.7 Max $2.77 $8.31 $3.21 sibling 1M Current Max flagship
§ 06 / DEEP LINKS

Specific scenarios.

All calculators →

Audit links

Provider hub →

Frequently asked.

Short answers for teams comparing Qwen3 30B A3B Thinking 2507 against other current Qwen3 text models.

Q · 01 What is the Qwen3 30B A3B Thinking 2507 input price? +
Alibaba's Singapore/International pricing row lists 1.468 CNY per 1M input tokens, converted here to $0.2168/M at 1 CNY = $0.14768.
Q · 02 What is the Qwen3 30B A3B Thinking 2507 output price? +
Qwen3 30B A3B Thinking 2507 is shown at $2.6/M output on this page. The original vendor row is 17.614 CNY per 1M thinking-mode output tokens.
Q · 03 Does Alibaba list prompt-cache pricing for this row? +
No exact cache-read token price is listed in the verified row, so the calculator keeps cache disabled instead of inventing a discount.
Q · 04 Why does this page use International pricing? +
AI//COST uses the Singapore/International row for Alibaba pages because it is the relevant baseline for non-mainland deployment and differs from some US/EU global rows.
Q · 05 Is this a text LLM page only? +
Yes. This page covers the text/code API model. Audio, image, and video models are intentionally handled separately.
Q · 06 Where did the context window come from? +
Context comes from Qwen's official model/blog documentation: Qwen model card lists 262,144 native tokens and a 1M-token long-context option. Pricing comes separately from Alibaba Model Studio.
Y
Reviewed by Yaroslav Vikhariev Founder - AI//COST - Last verified Jun 09, 2026 - Source: help.aliyun.com
Methodology Report a correction More by Y.V.