A small reference of published token prices. Filter the visible routes, then estimate a scenario from input and output volume—without claiming a universal best deal.
9 published price records · USD per million tokens; estimate excludes taxes, caching, storage, tools, batch pricing, minimum spend and discounts. This is not an invoice: confirm the current price before paying.
Alibaba Cloud Model Studio
Qwen3.8 Flash
qwen3.8-flash · Singapore International
Estimate$6.20
Input / 1M
$0.150
Output / 1M
$0.470
Plan and availability
International API price · Singapore International region
Limits and assumptions
Regional and plan prices can differ.
Z.AI
GLM-5.3-Flash
glm-5.3-flash · Published standard price
Estimate$6.50
Input / 1M
$0.150
Output / 1M
$0.500
Plan and availability
Global USD API pricing · Z.AI API
Limits and assumptions
Does not include Coding Plan pricing.
DeepSeek
DeepSeek V4.1 Flash
deepseek-flash · Off-peak hours
Estimate$7.50
Input / 1M
$0.150
Output / 1M
$0.600
Plan and availability
Standard API price; uncached input · DeepSeek API route
Limits and assumptions
All other hours are listed as off-peak; Chinese-holiday treatment follows the provider schedule.
DeepSeek
DeepSeek V4.1 Flash
deepseek-flash · Peak hours
Estimate$15.00
Input / 1M
$0.300
Output / 1M
$1.20
Plan and availability
Standard API price; uncached input · DeepSeek API route
Limits and assumptions
Peak schedule: UTC weekdays 01:00–04:00 and 06:00–10:00, except Chinese holidays.
DeepSeek
DeepSeek V4 Pro
deepseek-v4-pro · Off-peak hours
Estimate$26.40
Input / 1M
$0.660
Output / 1M
$1.98
Plan and availability
Standard API price; uncached input · DeepSeek API route
Limits and assumptions
All other hours are listed as off-peak; Chinese-holiday treatment follows the provider schedule.
DeepInfra
GLM-5.2
zai-org/GLM-5.2 · Standard list price
Estimate$31.50
Input / 1M
$0.750
Output / 1M
$2.40
Plan and availability
Base list price; not the temporary promotional quote · DeepInfra hosted inference route
Limits and assumptions
The page also showed a 25% promotional quote ($0.563 input / $1.80 output per million) without a published end date; it is not used in this calculator. This standard tier is fp4; quality or version equivalence to other GLM-5.2 routes is not tested here.
DeepSeek
DeepSeek V4 Pro
deepseek-v4-pro · Peak hours
Estimate$52.80
Input / 1M
$1.32
Output / 1M
$3.96
Plan and availability
Standard API price; uncached input · DeepSeek API route
Limits and assumptions
Peak schedule: UTC weekdays 01:00–04:00 and 06:00–10:00, except Chinese holidays.
Z.AI
GLM-5.2
glm-5.2 · Published standard price
Estimate$58.00
Input / 1M
$1.40
Output / 1M
$4.40
Plan and availability
Global USD API pricing · Z.AI API
Limits and assumptions
Does not include Coding Plan pricing.
Alibaba Cloud Model Studio
Qwen3.8 Max
qwen3.8-max · Singapore International · input ≤1M tokens
Estimate$80.00
Input / 1M
$2.00
Output / 1M
$6.00
Plan and availability
International API price · Singapore International region
Limits and assumptions
This row only covers the published input tier up to 1M tokens; regional and plan prices can differ.
Method
A calculator for the rows you can inspect.
Estimated cost = input millions × input price + output millions × output price. It excludes taxes, cached-token pricing, storage, tools, batch pricing, minimum spend, credits and discounts.
Time and region are part of the price.
DeepSeek entries are separate peak and off-peak scenarios; the published schedule uses UTC weekdays and has a Chinese-holiday exception. Alibaba entries are limited to the stated Singapore International scenario. Prices are not currency-converted on this page.
Scope and questions
Is this a full market comparison?
No. It is a deliberately small, dated reference of rows with an inspectable publisher link. It does not rank quality, speed or availability.
Why do two rows show one model?
They can represent different time or billing scenarios. They are alternatives, not separate providers or a recommendation.
Does the calculator send my estimates or require an API key?
No. There is no key input or estimate-submission endpoint. Calculations stay in this page's browser state; optional page analytics follow your existing consent settings.