Home / DeepSeek
A strong reasoning chat assistant that is free to use — the paid part is a separate per-token API, not a chat subscription
The free reasoning assistant — $0 for chat with a 1M context — the catch is it is a chat window, not a workspace, and the API is a separate meter.
DeepSeek chat is a free, strong reasoning assistant with a 1M-token context and no paid chat tier — the paid DeepSeek product is a separate per-token API for builders.
You want a capable second assistant for reasoning and long material at $0, and your workflow can live in a chat window without projects, scheduling or connectors.
You need the workspace layer the $20 assistants sell — projects, memory, scheduled tasks, custom agents and office connectors are not published features of the free chat product.
Workflows are built only from verified capabilities in our listing research — no vendor copy, nothing invented.
Starts A contract, report or codebase is too long for a normal chat window
You get Answers grounded in the full document, not an extract
Long-material work stops needing chunking and re-uploading
Starts A pricing, planning or technical problem needs slower reasoning
You get A reasoned answer you can interrogate
Approval Check any consequential recommendation independently before acting on it
A second, free opinion on the decisions that cost money
Starts A chat habit turns into a tool or automation you want to run programmatically
You get A programmatic version of the assistant you already know
The step from chatting to building does not require changing vendors
Take this week's hardest reasoning question or longest document and judge the answer, not a demo prompt.
Projects, files and scheduling stay in your existing tools — DeepSeek is the reasoning step, not the workspace.
If you later automate, read the official per-token table first: peak hours cost double the off-peak rate.
Time to first useful result: Under 1 hour · no-code 5/5
The chat product is free ($0) with no paid chat tier published. The separate DeepSeek API is billed per million tokens: Flash $0.15 input (cache miss) / $0.60 output off-peak, doubling at peak; V4-Pro $0.66 input / $1.98 output off-peak. Search and Fetch APIs are $0.75 per 1,000 successful requests.
$0 — free chat at chat.deepseek.com and in the app; no paid chat tier published
Flat / month · $0/mo for one person
| Plan | Published price | Solo monthly | Billed by | Limits & overage |
|---|---|---|---|---|
| ChatSolo pick | $0 — free chat at chat.deepseek.com and in the app; no paid chat tier published | $0/mo | Flat / month | No published projects, memory, scheduling or connector features for the chat product · Chat usage terms and any fair-use limits are the vendor's, not a published quota |
There is no chat subscription to jump between: the chat stays $0. The only spend appears if you choose to build on the API, where cost is tokens times the published per-million rate, with peak hours (01:00-04:00 and 06:00-10:00 UTC, Monday-Friday) billed at double the off-peak rate.
Price record · 2026-11-07: Checked 2026-11-07: the vendor's homepage (deepseek.com) routes users to free chat at chat.deepseek.com and to the API platform as two separate doors, and no paid chat tier is published on the official site — the chat product is modelled here honestly as a free listing at $0. The API meter was read directly this run on the official Models & Pricing page (api-docs.deepseek.com/quick_start/pricing/): prices per 1M tokens, off-peak rates half of peak — Flash input $0.003 cache hit / $0.15 cache miss and output $0.60 off-peak ($0.006 / $0.30 / $1.20 at peak); V4-Pro input $0.022 / $0.66 and output $1.98 off-peak ($0.044 / $1.32 / $3.96 at peak); Search and Fetch at $0.75 per 1,000 successful requests. Peak hours are 01:00-04:00 and 06:00-10:00 UTC Monday-Friday excluding Chinese public holidays, per the same page.
The same fields on every deep listing in this job, so agents compare field by field. “Not verified” means our research has not confirmed it yet — it never means “doesn't have it.”
DeepSeek's chat product — the assistant a solo founder actually opens at chat.deepseek.com and in the app — is free, with no paid chat tier published, as of November 2026; the money in DeepSeek is a separate, per-token API, stated plainly below. For everyday founder work that shape is unusually simple: you ask questions, work through documents and long material in a 1M-token context, switch thinking mode on for harder reasoning, and pay nothing for the chat itself. The current models behind it are DeepSeek-V4.1-Flash in the chat product's Instant and Expert modes and the V4-Pro model on the API, per the vendor's own docs. What you do not get is the surrounding product the $20 assistants sell: there is no published projects-and-memory workspace story, no scheduled agent tasks, no custom-agent marketplace and no bundled office connectors in the chat listing — it is a strong, free model in a chat window, and your workflow has to live around it. The API is a different purchase for a different moment: if you later build DeepSeek into a tool or an automation, the official pricing is per million tokens, split by peak and off-peak, cache hit and miss — Flash at $0.15 per million input tokens (cache miss, off-peak) and $0.60 per million output tokens off-peak, doubling at peak, with the Pro model at $0.66 and $1.98 off-peak. Budget the chat at $0, and budget any building separately.
Solo Fit stays the public score: the qualitative roll-up of these four published inputs under the How We Rank method. A genuinely free chat product and a five-minute start carry it; the trade is a thin workspace layer, recorded honestly in the checklist.
The difference: The full workspace assistant — projects, custom GPTs, scheduled tasks and Codex — the catch is the realistic plan costs $20/month where DeepSeek chat is $0.
The difference: The long-document writing specialist with Projects and connectors at $17-$20/month — the catch is a paid subscription with metered usage windows.
The difference: The assistant woven into Google's apps and search — the catch is its best work depends on living inside the Google ecosystem.
The chat product at chat.deepseek.com and in the app is free, and no paid chat tier is published on the official site, so the directory's True Solo Cost is $0 low and $0 typical. Money only enters if you separately choose to build on the API.
Per million tokens, off-peak: the Flash model charges $0.15 for input on a cache miss ($0.003 on a cache hit) and $0.60 for output; the V4-Pro model charges $0.66 for input on a cache miss and $1.98 for output. Peak hours — 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday — are billed at double those rates, per the official pricing page.
No paid chat plan is published. DeepSeek's official site offers the chat for free and prices only the developer API per token, which is why this listing is modelled as a free product with the API shown as an add-on, not as a subscription tier.
The workspace layer. No projects, scheduled tasks, custom-agent marketplace or bundled office connectors are published for the chat product, so organisation and automation stay in your other tools while DeepSeek does the reasoning step for $0.
DeepSeek is not in a published stack yet — stacks are edited buying guides, and memberships are added only when the combination has been worked through.
How we rank → · Get a weekly shortlist →