The free web version and API pricing are two separate systems. The API cost is very low but requires development skills. The web version's quota policy lacks official detailed explanation.
They also argued about these things on another page.
2026-09-21Daily auto-check logged — see the Chinese version for details
2026-09-20Daily auto-check logged — see the Chinese version for details
Show the remaining verification history (40 more)▼
2026-09-19Daily auto-check logged — see the Chinese version for details
2026-09-18Daily auto-check logged — see the Chinese version for details
2026-09-17Daily auto-check logged — see the Chinese version for details
2026-09-16Daily auto-check logged — see the Chinese version for details
2026-09-15Daily auto-check logged — see the Chinese version for details
2026-09-14Daily auto-check logged — see the Chinese version for details
2026-09-13Daily auto-check logged — see the Chinese version for details
2026-09-12Daily auto-check logged — see the Chinese version for details
2026-09-11Daily auto-check logged — see the Chinese version for details
2026-09-10Daily auto-check logged — see the Chinese version for details
2026-09-09Daily auto-check logged — see the Chinese version for details
2026-09-08Daily auto-check logged — see the Chinese version for details
2026-09-05Daily auto-check logged — see the Chinese version for details
2026-08-27Daily auto-check logged — see the Chinese version for details
2026-08-22Two rounds of checking (api-docs.deepseek.com/quick_start/pricing direct reading, consistent after reloading): ⚠️Confirmed price increase has taken effect - V4-Flash's cache hit for every million token inputs increased from US$0.0028 to US$0.007 (off-peak)/0.014 (peak), cache miss from 0.14 to 0.22/0.44, and output from 0.28 to 0.66/1.32, which is the implementation of the 'price adjustment plan' announced on the official website on 2026-08-11. At the same time, the official website added the deepseek-v4-pro model (which was not previously recorded), with a price approximately three times that of Flash. The model version has also been updated to DeepSeek-V4-Flash-0731/DeepSeek-V4-Pro-0813.
2026-08-14Local Claude parallel subagent recheck (read directly from api-docs.deepseek.com; current rates unchanged; the official site also announces that from 2026-08-16 it will switch to peak/off-peak dynamic pricing — the site's existing note already mentions "a price increase is coming," this is a clearer effective date and mechanism supplement)
2026-08-09 05:30:42Daily auto-check logged — see the Chinese version for details
2026-08-07 21:50:002026-08-07 21:50 Checked: The pricing of DeepSeek API is based on units of millions of tokens, divided into input and output categories. The prices for DeepSeek-V4-Flash and DeepSeek-V4-Pro are US$0.0028 and US$0.003625 for input (cache hit), US$0.14 and US$0.435 for input (miss), and US$0.28 and US$0.87 for output, respectively. The company expects to raise prices in the near future. Source: https://api-docs.deepseek.com/quick_start/pricing
2026-08-06 12:03:13Daily auto-check logged — see the Chinese version for details
2026-08-05 22:29:20Daily auto-check logged — see the Chinese version for details
2026-08-03 03:34:37Daily auto-check logged — see the Chinese version for details
2026-08-03(The official website api-docs.deepseek.com/quick_start/pricing was accessed successfully through the r.jina.ai agent) The six token pricing numbers for V4-Flash/V4-Pro remain unchanged; the official website has added an announcement that "peak hours (Beijing time 9:00-12:00, 14:00-18:00) will soon adopt a 2x fee rate", which has not yet taken effect and is not a current price change, and will continue to be monitored.
2026-07-31Daily auto-check logged — see the Chinese version for details
2026-07-30Daily auto-check logged — see the Chinese version for details
2026-07-30Daily auto-check logged — see the Chinese version for details
2026-07-30Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-27Daily auto-check logged — see the Chinese version for details
2026-07-26 11:02:57Daily auto-check logged — see the Chinese version for details
2026-07-26Daily auto-check logged — see the Chinese version for details
2026-07-26Daily auto-check logged — see the Chinese version for details
2026-07-26Daily auto-check logged — see the Chinese version for details
2026-07-25Daily auto-check logged — see the Chinese version for details
2026-07-25Daily auto-check logged — see the Chinese version for details
2026-07-25Daily auto-check logged — see the Chinese version for details
Continuously re-checked, every fact dated
Has a free plan
What Is This
Chinese open-source model family; web version free, API extremely cheap and widely popular.
Free Version Limitations
The web version is basically free, but has peak traffic limits.
Available Models And Quantity
Model
Free quota
Calculate Time Window
Description
DeepSeek V4 Flash
The web version is basically free, but it is rate-limited during peak hours.
Rate Limiting
API requires additional payment
DeepSeek V4 Pro
The web version is basically free, but it is rate-limited during peak hours.
Rate Limiting
API requires additional payment
Free Usage Limit
Features
Free quota
Description
Dialogue and programming
Available (with peak traffic limiting)
Context length
Up to 1M tokens (API specification)
Maximum output
384K tokens (API specification)
Quota Resets At
Peak hours have real-time rate limiting, with no clear daily reset
Terms Of Use
The web version can be used without registration, API requires payment, available for users in Taiwan, no regional restrictions, and supports Chinese and English interfaces
Best for: Developers, AI application integrators, and enterprises that require cheap API costs
API pay-as-you-go (usage-based)
Per million tokens (off-peak/peak): cache hit input US$0.007/US$0.014, cache miss input US$0.22/US$0.44, output US$0.66/US$1.32. Peak hours are 01:00-04:00 and 06:00-10:00 UTC, with peak prices being twice the off-peak prices. ⚠️ The price increase has taken effect (originally US$0.0028/US$0.14/US$0.28, this confirmation is the implementation of the previously announced 'planned price adjustment' on the official website, checked twice on 2026-08-22 with consistent results)
Which Model
DeepSeek V4 Flash
Usage Quota
Billed based on actual token consumption, with no monthly limit
Max Reading Time
Up to 1M tokens context, maximum output 384K tokens
This Plan Includes
Pay-as-you-go, pay for what you use
No contract or minimum consumption requirements
Supports large context (1M tokens)
Supports caching mechanism (reduces cost)
API documentation is complete, SDK supports multiple languages
Not Included
Web UI interface
Customer support (community-driven)
What Sets Us Apart
The world's cheapest LLM API (less than US$1 USD per million tokens). 50-90% cheaper than OpenAI/Claude API, suitable for cost-sensitive applications; the downside is that it is a Chinese company, API documentation is mainly in Simplified Chinese, and technical support is limited.
Best for: Developers, AI application integrators, and enterprises that require cheap API costs
API pay-as-you-go (usage-based)
New model (previously not recorded). Per million tokens (off-peak/peak): input cache hit US$0.022/0.044, input cache miss US$0.66/1.32, output US$1.98/3.96. Concurrent limit 500 (V4-Flash is 2500)
Which Model
DeepSeek V4 Pro
Usage Quota
Billed based on actual token consumption, with no monthly limit
Max Reading Time
Up to 1M tokens context, maximum output 384K tokens
This Plan Includes
Pay-as-you-go, pay for what you use
No contract or minimum consumption requirements
Supports large context (1M tokens)
Supports caching mechanism (reduces cost)
API documentation is complete, SDK supports multiple languages
Not Included
Web UI interface
Customer support (community-driven)
What Sets Us Apart
The world's cheapest LLM API (less than US$1 USD per million tokens). 50-90% cheaper than OpenAI/Claude API, suitable for cost-sensitive applications; the downside is that it is a Chinese company, API documentation is mainly in Simplified Chinese, and technical support is limited.
The world's most widely used AI assistant, handling conversation, search, image generation, and voice all in one place
FreeNT$0/month, open to everyone. The official site lists the free-tier limits one by one: limited access to GPT-5.5 Instant, limited messages and uploads, limited and slower image generation, limited deep research, limited memory and context, limited Codex, limited ChatGPT Work desktop app. The only place the official site gives concrete numbers is the plan comparison table: free-tier GPT Instant total context window 27K (Go/Plus 54K, Pro 128K); free-tier single-input cap about 12 pages of text (Go/Plus about 40 pages, Pro about 250 pages); the context window for reasoning models is marked depends on the situation for the free tier (Go/Plus 256K, Pro 400K). Response time on the free tier is marked as limited by system resources and service conditions; only paid plans are fast. Chat history is unlimited on the free tier. Whether content is used for model training can be opted out of. The official site has never published a concrete messages-per-day figure, and this site does not fill in an estimate. (Read directly from chatgpt.com/zh-Hant/pricing over a Taiwan connection on 2026-08-09)
Long-form writing and document quality are its strengths, with the paid version including the Claude Code engineering tool
FreeThe quota is calculated based on a rolling 5-hour usage window (not simply resetting at a fixed number every day), and general conversations can be used for around 10-20 times, excluding Claude Code.
Google has the deepest ecosystem integration of AI, with a generous free version and frequent discounts for student plans.
FreeNT$0/month, free to anyone with a Google account. Model available: Gemini 3.6 Flash; the official site states plainly that access to 3.1 Pro may vary, meaning free-tier access to Pro fluctuates and is not guaranteed. Features included on the free tier: image generation and editing, Deep Research, Gemini Live, Canvas, Gems, Gemini Notebook (research and writing), the Google Flow creative studio, plus limited usage of Nano Banana Pro. Cloud storage 15 GB (shared across Gmail/Drive/Photos). How the quota is counted and when it resets, in the official footnote's own words: the Gemini app measures usage limits by compute, which depends on prompt complexity, the features you use and conversation length; usage resets every 5 hours, up to a weekly usage cap. In other words the free tier is not a fixed number of messages per day but a compute allowance that resets every 5 hours plus an overall weekly cap; long conversations, Deep Research and image generation burn through it faster than plain chat. When the quota runs out you can buy AI credits. The official site does not publish the specific compute figure for the free tier. (Read directly from gemini.google/subscriptions over a Taiwan connection on 2026-08-09)
Support for mobile devices (iOS/App/tablet) is still being verified and only the web version is listed for now. The official website (api-docs.deepseek.com) indicates that the old model names deepseek-chat/deepseek-reasoner have been deprecated since 2026/7/24 and are now replaced by the general/thinking modes of DeepSeek-V4-Flash, and API users using the old model names should be aware of this change. As of 2026-07-27, a new observation item has been added: multiple media outlets (SCMP/winbuzzer.com/thenextweb.com) have reported that DeepSeek will introduce a peak hour pricing mechanism for V4, doubling the fees during peak hours (9am-noon and 2-6pm Beijing time), but as of this review, the official pricing page still shows a single rate and no price increase clause has been found, indicating that this is an upcoming change that has not yet been implemented, and this site will continue to monitor the situation.
Related Guide：DeepSeek to Raise Prices from 8/17: Its Priciest Item Jumps 12x, and Taiwan's Whole Workday Falls in Peak Hours
DeepSeek implements peak-valley pricing starting midnight 8/17. V4 Pro peak output pricing rises from ¥6 to ¥27 per million tokens, and the cache-hit price becomes 12x its original. Peak hours are 9-12 and 14-18 Beijing time, which fully overlap with Taiwan. Includes the official price list and four things to do before 8/17. Last verified: 2026-08-14
Looking for DeepSeek deals, discounts or promo codes? We check the official site automatically every day, review changes by hand, and date-stamp every entry.
