DeepSeek Pricing (2026): Free Chat, API Under $1/M Tokens

DeepSeek chat costs $0. Its API is usage-billed at $0.14 input and $0.28 output per million tokens for V4 Flash, or $0.435 and $0.87 for V4 Pro. The call is simple: keep personal chat free, default production work to Flash, and pay for Pro only when it cuts retries enough to beat a 3.11x price multiple.

DeepSeek pricing at a glance

DeepSeek has one free consumer product and two paid API models. There is no monthly or annual self-serve API subscription: the API deducts token charges from your account balance as you use it.

These numbers were verified against DeepSeek's live Models & Pricing page on August 12, 2026.

A token is the unit the model bills, not a request or a word. Input is everything sent to the model, including instructions, history, documents, and tool context. Output is everything the model returns. A cache hit is repeated input that DeepSeek can reuse at the lower rate; a cache miss is input it must process at the full rate.

The two API models share a 1M-token context window and a 384K maximum output. That removes the usual need to pay a long-context surcharge inside DeepSeek. Their operating limits differ: V4 Flash lists 2,500 concurrent requests, while V4 Pro lists 500.

DeepSeek chat is free, but the API is a separate bill

DeepSeek's consumer app costs nothing. Its official app announcement says it is 100% free, with no ads and no in-app purchases, and the current homepage still describes chat access as free.

Free does not mean the API is free. The consumer app is an interface for a person; the API is the metered product software calls. DeepSeek's API page deducts usage from a granted or topped-up balance. A free chat account therefore does not turn programmatic traffic into free traffic.

The same product split catches buyers elsewhere: an app subscription and API billing can share a brand while remaining separate products, as the Perplexity pricing breakdown shows. For DeepSeek, the distinction is even cleaner because the app has no published subscription price while every API token has a line item.

Who never needs to pay

You never need to pay if your use is interactive, occasional, and satisfied by the official web or mobile app. That covers personal questions, ad hoc writing, document discussion, and students who do not need to connect DeepSeek to another product.

DeepSeek does not publish a student-specific plan or discount on the pricing, FAQ, or app pages verified for this review. It does not need one for ordinary consumer access because that access already costs $0.

The boundary is automation. The moment a CRM, support workflow, internal tool, or customer-facing product needs to call DeepSeek without a person opening the app, budget for API tokens. Also avoid treating free chat as production infrastructure: the free-app page does not promise unlimited capacity or publish a service level.

V4 Flash is the production default; V4 Pro is an escalation lane

DeepSeek V4 Flash-0731 should receive the first request in most production routes. It has the lower price, the higher listed concurrency limit, and the broader current integration surface.

V4 Flash-0731

V4 Flash charges $0.14 per million cache-miss input tokens, $0.0028 for cache-hit input, and $0.28 for output. It supports thinking and non-thinking modes, JSON output, tool calls, DeepSeek's Anthropic-compatible API, and the Responses API. The live page lists a 1M context window, 384K maximum output, and 2,500 concurrent requests.

That combination makes Flash the sensible route for classification, extraction, summarization, first-draft generation, routine code changes, and bounded agent steps. These jobs have clear acceptance checks, so a failed output can be detected and escalated without letting a cheap mistake reach a customer.

Skip Flash as the only route when failure is expensive and hard to detect. A legal, financial, or customer-impacting recommendation should not choose its model from token price alone. The low rate buys room to evaluate; it does not replace evaluation.

V4 Pro

V4 Pro charges $0.435 per million cache-miss input tokens, $0.003625 for cache-hit input, and $0.87 for output. It has the same 1M context window and 384K maximum output, plus the same thinking modes, JSON output, tool calls, and Anthropic-compatible format. Its listed concurrency limit is 500.

Pro is the hard-tail route: the ambiguous request, the complex plan, the difficult coding task, or the second attempt after Flash fails a deterministic check. It is not the default merely because its name says Pro.

There is also an integration limit hidden inside the live table. DeepSeek currently lists Responses API support for Flash but not Pro. The page still says Pro support was planned for early August 2026, yet the live status on August 12 remains unsupported. If your application depends on that endpoint, Pro's token price is irrelevant until the integration exists.

DeepSeek's token bill is tiny; cost per accepted result still decides

The raw API expense is small enough that one avoidable human review can outweigh thousands of model calls. The useful formula is not cost per request. It is cost per accepted result.

For a request with no cache hits:

cost = input tokens / 1,000,000 × input rate + output tokens / 1,000,000 × output rate

At 1,000 tokens, V4 Flash costs $0.00014 for cache-miss input and $0.00028 for output. V4 Pro costs $0.000435 for cache-miss input and $0.00087 for output.

Consider a support-drafting workflow with 2,000 input tokens and 500 output tokens per reply. At 1,000 replies, Flash costs $0.42 and Pro costs $1.305 with no cache hits. The entire Pro premium is $0.885 per thousand replies.

Now add labor. Five minutes of review at an illustrative $60 hourly cost is $5. One extra review event costs more than five full batches of that Pro premium. The right question is therefore not, "Which model has the lower line item?" It is, "Which route clears the acceptance check with the least model spend plus review time?"

What a monthly DeepSeek bill looks like

DeepSeek has no seat-month to compare with a software subscription. Monthly cost is simply the traffic shape: how much context goes in, how much text comes out, how often prefixes hit cache, and how many attempts an accepted result takes.

For a solo product prototype using 10M cache-miss input tokens and 2M output tokens in a month, Flash costs $1.96 and Pro costs $6.09. The $4.13 difference will not decide whether the product survives. The model that produces a valid result with less debugging will.

For a support operation drafting 100,000 replies at 2,000 input and 500 output tokens each, the uncached bill is $42.00 on Flash or $130.50 on Pro. At an illustrative 80% input cache-hit rate, those totals fall to $20.048 and $61.48. Even at this volume, one hour of human review can exceed the entire monthly Flash bill under the $60-per-hour assumption.

For 1,000 document-analysis jobs using 50,000 input and 2,000 output tokens each, the uncached bill is $7.56 on Flash or $23.49 on Pro. With 80% of input hitting cache, it becomes $2.072 or $6.235. Long input does not automatically mean a large DeepSeek bill; repeated long input can make the unit economics unusually small.

These examples also show why a fixed monthly estimate copied from another company is useless. A support reply is output-heavy relative to its prompt. A document job is input-heavy. An agent can multiply both through retries and tool loops. Save the token mix beside the dollar figure so finance can tell whether traffic, quality, or a vendor rate caused the change.

The Flash-to-Pro break-even

V4 Pro's cache-miss input and output rates are both 3.107143 times Flash, rounded to 3.11x. Under the same token mix, Pro must reduce billed attempts or token volume by 67.82% to break even on model spend alone.

If Pro finishes a task in one attempt, Flash can average up to 3.11 comparable attempts before its token bill becomes higher. If Flash succeeds in one or two attempts, Pro loses the price comparison. If Flash repeatedly needs four attempts, Pro wins even before human review enters the equation.

This routing rule also prevents a common procurement mistake. A team sees rates below one dollar, sends every task to Pro, and calls the total cheap. The absolute bill may still be small, but an unmeasured premium across millions of calls is not a strategy. Flash should own the bounded base load; Pro should own the failures that justify it.

Context caching changes the bill only when prefixes truly repeat

DeepSeek's cache-hit rate is valuable, but it is not a discount you can assume in a spreadsheet. The context-caching guide says caching is enabled by default, works on complete persisted prefix units, and operates on a best-effort basis.

A stable system prompt followed by a stable reference document is a strong cache candidate. A later request can reuse that beginning and append a new question. Reusing a phrase from the middle of a different prompt does not qualify. The complete prefix must match a unit DeepSeek has persisted.

The cache also has a time dimension. Construction takes seconds, and unused entries are normally cleared within hours to days. A weekly report may miss a cache that a burst of near-identical requests would hit. DeepSeek does not guarantee a 100% hit rate.

At an illustrative 80% input cache-hit rate, the blended input cost per million tokens falls to $0.03024 on Flash and $0.08990 on Pro. In the 1,000-reply support example, total cost falls from $0.42 to $0.20048 on Flash and from $1.305 to $0.61480 on Pro.

That is roughly half the total bill in this output-bearing workload, not the near-total reduction suggested by comparing cache-hit and cache-miss input rates in isolation. Output tokens do not receive the input cache discount.

1. 1

   Keep the prefix stable

   Put durable instructions and repeated reference material first. Append the changing user request after that stable prefix instead of rebuilding the prompt in a different order.

2. 2

   Read the usage fields

   Track prompt_cache_hit_tokens and prompt_cache_miss_tokens from API responses. A projected hit rate is not a billing fact until these fields show it.

3. 3

   Price the observed mix

   Split measured input into hit and miss tokens, add output tokens, and calculate cost per accepted result. Reprice when prompt structure or traffic cadence changes.

DeepSeek is cheaper than the API shortlist, before quality adjustment

DeepSeek wins the raw-token comparison by a wide margin. It does not automatically win the business outcome because these models differ in capability, tools, latency, and failure rate.

Use one normalized workload to make the sticker prices comparable: 1M total input tokens plus 250K output tokens, with no cache discount. That workload costs $0.21 on V4 Flash and $0.6525, rounded to $0.65, on V4 Pro.

OpenAI GPT-5.6 Sol charges $5 input and $30 output per million tokens. The normalized workload costs $12.50 when each request stays below OpenAI's long-context threshold, or 19.16x V4 Pro. Prompts above 272K input tokens trigger 2x input and 1.5x output pricing for the full request, and cache writes cost 1.25x uncached input.

Claude Sonnet 5 charges $2 input and $10 output per million tokens. The same workload costs $4.50, or 6.90x V4 Pro. Anthropic includes the 1M context window at standard pricing and offers a 50% Batch discount, so latency-tolerant work can narrow the gap.

Gemini 3.1 Pro Preview charges $2 input and $12 output per million tokens when a prompt stays at or below 200K. The normalized aggregate workload costs $5.00 if every request stays under that boundary, or 7.66x V4 Pro. Put the same workload into one request above 200K and Google's rates rise to $4 input and $18 output, producing an $8.50 bill, or 13.03x V4 Pro.

The price spread is large, but the dollar spread is smaller than the multiples make it sound. Moving the normalized workload from V4 Pro to GPT-5.6 Sol adds $11.8475. One human correction can erase that saving, while a reliable high-volume route can compound it.

Use the DeepSeek vs ChatGPT comparison when model quality, product features, and data handling matter more than the token table. Use the cheapest AI API shortlist when the job can accept smaller budget models and gateways. A pricing page can normalize rates; only your evaluation can normalize outcomes.

The hidden budget risk is repricing, not an annual lock

DeepSeek's live page says an overall API price increase is planned in the near future and describes the expected increase as significant. It publishes neither an amount nor an effective date. Today's rates are verified facts; any future multiplier is speculation.

That warning changes the procurement decision. Do not promise a customer, board, or finance team that the August rate will hold for a year. Store the pricing-page verification date beside every budget model, make the rate an editable input, and define the price change that would trigger a reroute.

There is no monthly or annual API trap

DeepSeek lists token-based deduction, not a monthly seat or annual self-serve contract. Topped-up balance does not expire. Granted balance can expire, and its date appears on the Billing page. The API uses granted balance first when both balances exist.

Unused balance is refundable through Billing > Refunds, according to DeepSeek's official FAQ. That makes a modest top-up less risky than a non-refundable annual commitment. Account deletion is the exception: the FAQ warns that deleting the account forfeits any remaining Open Platform balance.

DeepSeek also avoids a postpaid overage surprise. When balance is insufficient, the API returns HTTP 402 and tells you to top up. The operational risk is interruption, not a bill that silently runs beyond a plan limit. Set a balance alert before production traffic finds that boundary for you.

Integration and concurrency can cost more than tokens

Flash lists 2,500 concurrent requests and current Responses API support. Pro lists 500 and no current Responses API support. A buyer who designs around Pro without checking those limits may pay in queueing, fallback logic, or integration work even while the model bill stays below one dollar per million tokens.

The cheapest safe architecture keeps routing reversible. Use one internal interface, record the chosen model and token usage, and preserve a tested fallback. DeepSeek's price warning makes that ordinary engineering discipline a budget control.

DeepSeek's price history split Flash from Pro

V4 did not make every line item uniformly cheaper. The current tier strategy pushed Flash down aggressively while preserving a higher cache-miss input price for Pro.

DeepSeek's V3 pricing announcement listed $0.07 cache-hit input, $0.27 cache-miss input, and $1.10 output per million tokens. Against those rates, V4 Flash is 96.00% cheaper on cache-hit input, 48.15% cheaper on cache-miss input, and 74.55% cheaper on output.

V4 Pro moves differently. Its cache-hit input is 94.82% cheaper than V3 and its output is 20.91% cheaper, but its $0.435 cache-miss input is 61.11% higher than V3's $0.27.

That history explains the current decision rule. Flash is not merely a smaller badge; it is the volume-pricing play. Pro asks the workload to earn a higher uncached rate through better outcomes. The next announced increase may change both, so the verification date belongs next to the arithmetic.

The Monday move: measure one route before moving the budget

On Monday, take 100 sanitized requests from one bounded workflow and run a one-week routing test. Do not start with an enterprise-wide migration.

1. 1

   Define acceptance first

   Choose a result a reviewer can score consistently: valid schema, correct classification, passing tests, or an approved support draft. Remove customer, personal, and confidential data from the evaluation set.

2. 2

   Send the base load to Flash

   Run all 100 requests through V4 Flash-0731. Record input, output, cache-hit, cache-miss, retry, latency, and acceptance data for each request.

3. 3

   Escalate only failures

   Send failed or ambiguous cases to V4 Pro. Measure whether Pro cuts the failed tail by enough to beat the 3.11x model rate and any human review time.

4. 4

   Set two budget triggers

   Keep Flash as the default only while it clears the quality floor. Keep DeepSeek as the provider only while the live normalized cost stays below your tested fallback after any price change.

The output of this test is one decision-ready number: cost per accepted result. It converts DeepSeek's tiny token rates into a workflow budget finance can revisit without rebuilding the analysis.

Frequently asked questions

Is DeepSeek paid or free?

DeepSeek's official consumer app is free, with no ads or in-app purchases. API access is separate and charges for cache-hit input, cache-miss input, and output tokens.

How much does DeepSeek V4 Pro cost?

V4 Pro costs $0.003625 per million cache-hit input tokens, $0.435 per million cache-miss input tokens, and $0.87 per million output tokens. It has a 1M context window, 384K maximum output, and a listed concurrency limit of 500.

How much is DeepSeek per month?

Consumer chat costs $0 per month. The API has no fixed monthly price: multiply monthly cache-hit input, cache-miss input, and output token totals by their per-million rates.

How is DeepSeek cheaper than ChatGPT?

On a price-only workload of 1M input plus 250K output tokens, DeepSeek V4 Pro costs about $0.65 and OpenAI GPT-5.6 Sol costs $12.50 below its long-context threshold, a 19.16x difference. Compare accepted outcomes, not tokens alone, because model quality and tool support differ.

Does DeepSeek offer a student discount?

DeepSeek does not list a student-specific plan or discount on the official pricing, FAQ, or app pages verified in August 2026. Students using the consumer web or mobile app already pay $0.

Can I get a refund from DeepSeek?

Yes, unused API balance is refundable through Billing > Refunds. Topped-up balance does not expire, while granted balance can carry an expiration date. Deleting the account forfeits remaining Open Platform balance.

Did DeepSeek pricing change in 2026?

Yes. Against the prior V3 schedule, V4 Flash is 48.15% cheaper on cache-miss input and 74.55% cheaper on output. V4 Pro output is 20.91% cheaper, but its cache-miss input is 61.11% higher. DeepSeek now warns that another significant increase is planned without stating the amount or date.

Should I choose V4 Flash or V4 Pro?

Choose Flash for the bounded base load and Pro for failed or high-consequence cases. On uncached model spend, Pro needs about 67.82% fewer billed attempts or tokens to break even with Flash's 3.11x lower rate.

Get the AI Tools Map for Business Owners

The AI Tools Map for Business Owners turns model prices into a practical adoption stack, with the quality floor, routing role, and cost trigger for each tool. Subscribe to get the next edition free.