The most expensive DeepSeek mistake is assuming free chat covers bots. As of the 17 September 2026 pricing page, consumer chat is free with no paid chat tier; the API is pay-per-token. Automations need keys, model IDs, and a rate card. Recheck platform.deepseek.com whenever you quote a client.
What free chat includes
Free consumer chat is for interactive Q&A in DeepSeek’s chat UI. Per the 17 September 2026 pricing page, there is no paid chat tier in that report. That is great for learning, drafting, and pasting code. It is a poor fit for CI bots, customer-facing assistants, or scheduled jobs. Those need the API. If DeepSeek changes chat monetization later, you will only know by rechecking the platform—not by trusting this paragraph forever. Treat every rate and limit as a dated snapshot from DeepSeek or the other vendor docs linked in sources, not as a permanent contract price.
What the API bills
API traffic uses deepseek-flash or deepseek-v4-pro with peak/off-peak and cache pricing from the 17 September 2026 page: Flash $0.30/$1.20 peak per million (off-peak half), Pro $1.32/$3.96 peak (off-peak half), plus much cheaper cached input. Specs include 1M context and 384K max output; Flash has vision and higher concurrency. Meter every environment. Free chat login cookies are not API auth. Treat every rate and limit as a dated snapshot from DeepSeek or the other vendor docs linked in sources, not as a permanent contract price. Recheck platform.deepseek.com for Flash and Pro cards, and confirm peak UTC
Mixing chat, API, and editors
A common India freelance stack: DeepSeek free chat for quick questions, DeepSeek API for a client bot, Cursor Start India for editing (Cursor models only at ₹649/mo). None of those fees cancel the others. Cursor Pro’s Other Models pool bills at API rates if you pull non-Cursor models there. Read Cursor docs and platform.deepseek.com so you do not double-count or under-count. Treat every rate and limit as a dated snapshot from DeepSeek or the other vendor docs linked in sources, not as a permanent contract price. Recheck platform.deepseek.com for Flash and Pro cards, and confirm
Related on this site
Frequently asked questions
-
Can my app use free chat under the hood?
No. Apps should use the API with keys. Free chat is the consumer product. Automating a chat UI scrapes a path you should not rely on for production. Confirm live details on the vendor docs linked in this brief before you quote a client or ship production traffic. Attribute DeepSeek Flash and Pro figures to the 17 September 2026 pricing page when you reuse them, then recheck platform.deepseek.com.
-
Is there a paid DeepSeek chat plan?
As of the 17 September 2026 pricing page, reporting says no paid chat tieru2014consumer chat is free. Recheck platform.deepseek.com for changes. Confirm live details on the vendor docs linked in this brief before you quote a client or ship production traffic. Attribute DeepSeek Flash and Pro figures to the 17 September 2026 pricing page when you reuse them, then recheck platform.deepseek.com.
-
Which API model is cheapest?
Flash has lower listed rates than Pro, especially off-peak and with cached input. Numbers are from 17 September 2026u2014confirm live before you promise unit costs. Confirm live details on the vendor docs linked in this brief before you quote a client or ship production traffic. Attribute DeepSeek Flash and Pro figures to the 17 September 2026 pricing page when you reuse them, then recheck platform.deepseek.com.
-
Does free chat have 1M context?
The 1M context and 384K max output figures in this brief are from the API model cards for Flash and Pro. Do not assume identical chat UI limitsu2014check the product you actually use. Confirm live details on the vendor docs linked in this brief before you quote a client or ship production traffic. Attribute DeepSeek Flash and Pro figures to the 17 September 2026 pricing page when you
Sources
Bdeb Technology builds websites, WordPress systems, and tools on top of models like these. A written quote comes back within 24 hours.