DeepSeek Pricing
Per-token API rates, cache-hit discounts, the free consumer tier, self-hosted GPU economics, and honest comparisons to OpenAI, Anthropic, and Google.
DeepSeek has become the benchmark for low-cost large-model inference. But buyers tend to ask the same three questions: What does the API actually cost per million tokens? Is the chat app genuinely free? And does that cost advantage hold up in production once support, compliance, and reliability enter the picture?
This page answers those questions using figures from DeepSeek's own pricing page, cross-checked against its current model lineup. Prices were verified on 2026-09-21. DeepSeek can change its rates, so check the DeepSeek API pricing page for the latest figures before committing to a contract.
A quick note on model names: since the 2026-09-10 pricing update, DeepSeek's live API models are deepseek-flash (DeepSeek-V4.1-Flash) for most workloads and deepseek-v4-pro for larger, higher-quality workloads. The older deepseek-v4-flash, deepseek-chat and deepseek-reasoner names route to the current lineup. If a pricing table still lists them separately, it is out of date.
We use DeepSeek in production for clients with high-volume, mostly English, internal workloads. That does not make it the right choice for every use case. Before migrating, read the comparison section carefully.
DeepSeek Pricing at a Glance
There are four ways to pay for DeepSeek, and they price very differently:
1. DeepSeek Chat (web and app) — free for individual consumer use. No paid subscription tier is currently advertised on deepseek.com.
2. DeepSeek API — per-token, priced in USD per 1 million tokens. Two tiers: deepseek-flash (V4.1-Flash) for most workloads and deepseek-v4-pro for higher-quality output. Both apply a large cache-hit discount on repeated input tokens.
3. Self-hosted via open weights — no per-token fee. You pay for GPU hardware or GPU rental, plus operations. Break-even against the API depends on utilization.
4. Third-party hosts — providers like Together, Fireworks, and DeepInfra host DeepSeek's open-weight models with their own price cards. Use these when you want an alternative jurisdiction or SLA.
The API prices below are what DeepSeek charges directly. Cache-hit input pricing is roughly 30x to 50x cheaper than cache-miss input pricing on both deepseek-flash and v4-pro, so caching design has more impact on your bill than model choice does.
Prices verified 2026-09-21 against DeepSeek's official developer documentation.
API Pricing per Model
DeepSeek publishes prices per 1 million tokens, split between input and output, with a separate discounted rate when the input prefix hits the API's prompt cache. The numbers below are the direct-from-vendor peak rates as of 2026-09-21; DeepSeek bills 50% less outside its stated peak window (01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday), so check the current window on the pricing page.
deepseek-flash (DeepSeek-V4.1-Flash, released 2026-09-10; the legacy deepseek-v4-flash and deepseek-v4-flash-vision-exp IDs now route here) — Input cache miss: $0.30 per 1M tokens peak, $0.15 off-peak. Input cache hit: $0.006 per 1M tokens peak, $0.003 off-peak. Output: $1.20 per 1M tokens peak, $0.60 off-peak. This is DeepSeek's primary workhorse model and the price most third-party comparisons quote.
deepseek-v4-pro (DeepSeek-V4-Pro-0813, the larger tier) — Input cache miss: $1.32 per 1M tokens peak, $0.66 off-peak. Input cache hit: $0.044 per 1M tokens peak, $0.022 off-peak. Output: $3.96 per 1M tokens peak, $1.98 off-peak. Use this when the workload actually needs stronger reasoning or code quality; do not default to it.
How to read the cache-hit line: DeepSeek automatically caches your prompt prefix. Any subsequent request that reuses that prefix (a common system prompt, a shared retrieval context, few-shot examples that do not change per call) pays the cache-hit rate on those tokens instead of the cache-miss rate. On a typical RAG or agent workload the effective input price sits somewhere between the two numbers, weighted by how much of your prompt is reused.
DeepSeek announced on 2026-09-10 that deepseek-v4-pro would route to V4.1-Flash from 2026-09-14, then reversed that in its change log and kept V4 Pro on unchanged billing, so pin the model ID you tested and watch the change log before relying on either tier.
Off-Peak Discount Window
DeepSeek cuts API rates by 50% during off-peak hours for both deepseek-flash and deepseek-v4-pro on its pricing schedule. Peak hours run from 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday through Friday. All remaining weekday hours and the entire weekend bill at the half-price off-peak rate.
Peak prices for deepseek-flash are $0.30 per million input tokens on a cache miss, $0.006 on a hit, and $1.20 for output. For deepseek-v4-pro, peak tokens cost $1.32 on an input miss, $0.044 on a hit, and $3.96 for output. Off-peak processing costs exactly half those figures, making batch jobs and asynchronous evaluations much cheaper when scheduled outside peak UTC blocks.
This rate card took effect at 04:00 UTC on 2026-09-10, but the schedule is subject to revision. DeepSeek previously used a Beijing-overnight window and later removed off-peak discounts in mid-2026 before introducing this split. Recheck DeepSeek's pricing documentation before building automated pipeline triggers around these specific operating hours.
DeepSeek Chat (Consumer) Pricing
The DeepSeek Chat product at chat.deepseek.com and in the iOS and Android apps is free to use for individual consumers. There is no paid subscription tier equivalent to ChatGPT Plus or Claude Pro on DeepSeek's own consumer surface as of 2026-09-21.
Free-tier trade-offs: there are no published usage quotas, but the service is rate-limited and can be degraded during heavy load. Some features (file upload, extended reasoning, image input) may be gated, throttled, or region-restricted depending on the app version. For a walkthrough of getting started, running it locally and the safety trade-offs, see How to Use DeepSeek.
Important buyer note: the free chat product runs on infrastructure operated from mainland China and is subject to Chinese data-handling law. It is fine for personal experimentation. It is not appropriate for regulated data, client-confidential material, or anything a compliance team would want to audit. For that, use the API under a proper contract, a third-party host in a jurisdiction you accept, or a self-hosted deployment. See our DeepSeek data privacy and security risks guide for the full analysis.
Self-Hosted / Open-Weights Cost
DeepSeek's open-weight releases carry no per-token fee. Your cost is hardware plus operations. That trades a variable API bill for a fixed capacity cost, which only makes sense above a certain utilization threshold.
Hardware ballpark: the full-size DeepSeek V3 and R1 open-weight checkpoints are 671B-parameter Mixture-of-Experts models. Serving them at full precision requires multiple high-end datacenter GPUs (H100 or H200 class) in a single node — practical deployments typically use FP8 or INT8 quantization to fit onto a single 8-GPU H100/H200 box. Smaller distilled variants (7B to 70B parameters) run comfortably on one or two consumer or workstation GPUs.
Cloud-rental math: an 8x H100 node on the major GPU marketplaces rents for roughly $16-30 per hour depending on provider, commitment, and region. Run it 24/7 and you are looking at approximately $12,000-22,000 per month before egress, storage, and orchestration. Divide that by the tokens you actually generate to get your effective per-million-token cost.
When self-hosting beats the API: batch workloads, high sustained throughput, strict data-residency requirements, or a need to fine-tune on private data. When it loses: bursty traffic, low utilization, small teams, or any workload where you would end up under-using the hardware. For most teams under six figures of monthly API spend, the DeepSeek API is cheaper and less operational trouble than self-hosting.
See DeepSeek V3 Explained for the architecture and deployment detail behind these numbers.
DeepSeek Pricing on AWS Bedrock and Azure
You do not have to call DeepSeek's own API to use it. Both Amazon Web Services and Microsoft Azure list DeepSeek's open-weight models in their managed model catalogs, and a growing share of the searches for "DeepSeek pricing" are really asking what it costs through one of those two.
Amazon Bedrock offers DeepSeek models through the Bedrock Marketplace, billed the same way as Bedrock's other pay-per-use models: per-token, on your existing AWS bill, with no separate DeepSeek account or subscription. Microsoft Azure lists DeepSeek in the Azure AI Foundry model catalog, billed through your Azure subscription on the same consumption basis.
The tradeoff is consistent across both: you get AWS or Azure's SLA, IAM/RBAC, VPC networking, and procurement paperwork already in place, which matters for regulated or enterprise buyers. What you give up is DeepSeek's own cheapest cache-hit pricing — Bedrock and Azure meter usage on their own rate cards, which have historically run higher than DeepSeek's direct API, especially on cached input tokens.
Rule of thumb: call DeepSeek's API directly when price per token is the deciding factor. Use Bedrock or Azure AI Foundry when you need the request to stay inside an existing AWS or Azure compliance boundary and the marketplace premium is worth it for that. Current per-token rates on either platform change independently of DeepSeek's own pricing page, so check AWS Bedrock pricing or the Azure AI Foundry model catalog directly before you budget.
How DeepSeek Pricing Compares
Every figure in this section comes from the vendor's own public pricing page linked in sources. Vendor pricing changes frequently — treat these as the numbers we verified on 2026-09-21 and re-check before any procurement decision.
DeepSeek deepseek-flash (V4.1-Flash, verified 2026-09-21): $0.30 input cache miss / $1.20 output per 1M tokens at peak, half that off-peak.
OpenAI GPT-4o (as of 2026-07-21): approximately $2.50 input / $10.00 output per 1M tokens. GPT-4o mini: approximately $0.15 input / $0.60 output per 1M tokens. o1 and o3-class reasoning models are meaningfully more expensive on the output side. Check the OpenAI pricing page for the current live table.
Anthropic Claude Sonnet class (as of 2026-07-21): approximately $3.00 input / $15.00 output per 1M tokens. Haiku class is meaningfully cheaper. Check the Anthropic pricing page for the current lineup.
Google Gemini Pro (as of 2026-07-21): typically low single-digit dollars per 1M input tokens with tiered pricing above 128K context. Gemini Flash sits in the sub-dollar tier and is the closest direct competitor to deepseek-flash on price. Check the Google AI pricing page for the current split.
Straight read: DeepSeek deepseek-flash is meaningfully cheaper than GPT-4o and Claude Sonnet on a straight per-token basis, and about twice GPT-4o mini at peak rates, level with it off-peak. The gap widens further with cache hits, where DeepSeek's $0.006 per 1M input tokens (peak) is among the lowest published rates in the market. It narrows or disappears once you compare against the frontier vendors' cheapest tiers on tasks their small models can actually handle. For the head-to-head, see DeepSeek vs OpenAI and DeepSeek vs Claude.
Total Cost of Ownership for Business
Per-token pricing is not your bill. Your bill is per-token pricing multiplied by tokens per user session multiplied by users, plus caching design, plus support and compliance overhead.
Illustrative worked example — internal knowledge-base assistant, 100 employees, average 30 assisted sessions per person per week, average 4,000 input tokens (system prompt plus retrieved context) and 500 output tokens per session, 70% cache hit rate on the shared system prompt. That is roughly 12,000 sessions per week and about 52,000 sessions per month. On deepseek-flash, monthly token cost lands near $50 a month at peak rates and about $25 off-peak (145.6M cache-hit input tokens at $0.006, 62.4M cache-miss input tokens at $0.30, 26M output tokens at $1.20) depending on exact cache-hit ratio.
The same workload on GPT-4o at $2.50 input and $10 output lands in the low hundreds of dollars per month. On Claude Sonnet at $3 and $15, higher again. On GPT-4o mini or Gemini Flash, it lands close to DeepSeek. The dollar delta at this scale is a rounding error on any real IT budget — do not migrate for it.
Where DeepSeek economics start to matter is high-volume workloads: document processing at scale, agentic loops that consume tens of millions of tokens per day, background summarization jobs, or customer-facing chat with heavy usage. At that scale the per-token gap compounds into real money, and the case for either the DeepSeek API or a self-hosted deployment strengthens.
Costs the sticker price does not show: engineering time to swap providers, evaluation cycles to confirm quality parity on your tasks, monitoring and observability for a less-mature vendor, procurement and legal review of a non-Western vendor, and the ongoing cost of maintaining a second inference path for fallback. Model those in before you decide.
Is DeepSeek Really Cheaper?
On a straight per-token basis: yes, clearly. deepseek-flash is roughly one order of magnitude cheaper than Claude Sonnet and GPT-4o on output tokens, and the cache-hit input price is among the lowest published anywhere.
On a total-value basis: it depends on the workload. Frontier models still lead on complex multi-step reasoning, tool-use reliability, long-context recall past 128K, and multimodal tasks. If you swap a model that ships the job in one call for one that needs three retries, the cheaper per-token price evaporates.
Costs to weigh honestly against the per-token savings: (1) enterprise support — DeepSeek does not offer the same 24/7 incident response, dedicated account teams, or contractual SLAs that OpenAI, Anthropic, and Google sell to enterprise customers; (2) compliance — data-residency, GDPR, HIPAA, FedRAMP, and SOC 2 posture at DeepSeek is less mature and less documented than at the Western frontier labs, which matters in regulated industries; (3) reliability — third-party status pages have shown DeepSeek's API subject to capacity constraints during peak demand; (4) geopolitics — a growing number of Western buyers have policy restrictions on Chinese-hosted inference, which forces you into a third-party host or self-hosted path that erodes some of the price advantage.
Rule of thumb we apply for clients: DeepSeek wins on price when the workload is internal, high-volume, English, and the failure mode of an occasional wrong answer is acceptable. Frontier vendors win on price-plus-risk when the workload is customer-facing, regulated, low-volume, or reasoning-heavy. Most real businesses end up with both. If you are weighing the buy decision itself, Is DeepSeek-V4 Worth It? breaks it down by solo operator, small team and larger organization.
How to Choose the Right DeepSeek Plan
Personal experimentation, non-sensitive prompts, no budget line: use DeepSeek Chat on the web or app. Free, no signup friction, zero commitment.
Team pilot or production API workload under a few hundred million tokens per month, English text and code use cases, internal-facing: deepseek-flash via the DeepSeek API. Start with a strict monthly spend cap in your dashboard.
Higher-quality reasoning, complex code generation, or workloads where deepseek-flash quality is measurably insufficient on your evals: deepseek-v4-pro. Do not pay for pro until you have proved flash is not enough on your actual tasks.
Regulated industry, EU data residency, or a customer contract that forbids Chinese-hosted inference: skip the DeepSeek API entirely. Either self-host the open weights on infrastructure you control, or use a Western-jurisdiction third-party host that carries DeepSeek's open-weight models under their own compliance program. DeepSeek Alternatives covers Western-hosted and enterprise-supported options.
Sustained high throughput above roughly 500M-1B tokens per month with predictable load: model both the API bill and a self-hosted deployment. The crossover point moves with GPU market prices, but self-hosting can become cheaper at sustained enterprise scale.
Whichever path you pick, wire spend alerts in the DeepSeek console (or your GPU-cloud dashboard) before the first production call. That is the single control that has saved every client from a runaway prompt loop.
What you need to run DeepSeek yourself
DeepSeek is a frontier-scale Mixture-of-Experts model, so "running it yourself" is a real infrastructure decision — not something a single laptop or gaming GPU can do. Match the path below to how seriously you need to self-host. For most teams the API or rented GPUs are the right answer; buying hardware only pays off at steady, high volume or when your data can never leave your walls.
| Path | What it is | Best for | Get started |
|---|---|---|---|
| Call the hosted API | Use DeepSeek as a pay-per-token API — zero hardware | Most teams; evaluating before committing | OpenRouter |
| Rent GPUs by the hour | Spin up H100 / A100 nodes on demand, tear them down after | Self-hosting without capital outlay; bursty workloads | RunPod |
| Local on unified memory | A single workstation with enough unified memory to hold a 4-bit quant | One powerful on-prem box; privacy-first solo/SMB use | Apple Mac Studio (M3 Ultra, 512GB) |
| Local on workstation GPUs | Multiple 48GB professional cards for MoE offload / tensor parallelism | Power users and small clusters that want cards they own | NVIDIA RTX 6000 Ada (48GB) |
Once DeepSeek is running, the fastest way to put it to work day to day is inside Cursor — point it at the model through OpenRouter as a custom model. And if you would rather run a model on one affordable box, see Best mini PCs for local AI and Local AI hardware calculator.
Frequently Asked Questions
- DeepSeek Chat on the web and mobile app is free for consumers. The DeepSeek API charges per token: as of 2026-09-21, deepseek-flash (V4.1-Flash) is $0.30 per 1M input tokens (cache miss), $0.006 per 1M input tokens (cache hit), and $1.20 per 1M output tokens at peak, with all three halved off-peak. deepseek-v4-pro is roughly 3x to 7x more expensive on those lines. Self-hosting the open-weight models has no per-token fee but requires GPU hardware.
- The DeepSeek consumer chat product at chat.deepseek.com and in the iOS and Android apps is free to use, with no advertised paid subscription tier. The DeepSeek API is not free — it is paid per token, but the free tier consumer chat is functionally free for individual, non-sensitive use. Do not use the free chat product for regulated data or client-confidential material.
- Yes, meaningfully, when compared against GPT-4o class models. deepseek-flash at $0.30 input and $1.20 output per 1M tokens (peak) is roughly eight times cheaper than GPT-4o's approximately $2.50 input and $10 output per 1M tokens (verified 2026-09-21). Against GPT-4o mini at approximately $0.15 input and $0.60 output, DeepSeek is about twice the price at peak and level off-peak, so the saving against small models comes from cache hits, not the sticker rate. Always verify at the vendor pricing pages before contracting.
- Three real reasons: efficient Mixture-of-Experts architecture that activates only a fraction of parameters per token, aggressive engineering on inference cost, and a business model that appears to prioritize adoption and geopolitical positioning over margin. Buyers should also weigh what is not priced into the sticker: enterprise support, formal SLAs, Western compliance certifications, and jurisdictional risk are less mature than at OpenAI, Anthropic, or Google.
- Yes, on the API. DeepSeek prices in USD per 1 million tokens, with input tokens split between a cache-miss rate and a much lower cache-hit rate (roughly 50x cheaper on cache hits), and output tokens priced separately. The consumer chat product does not charge per token — it is free. Self-hosted deployment of the open-weight models has no per-token fee at all.
- No. As of 2026-09-21, DeepSeek does not advertise a consumer subscription tier equivalent to ChatGPT Plus or Claude Pro. The consumer chat product is free, and paid usage is exclusively via the metered API. If your workload needs guaranteed capacity or a support contract, that is negotiated as an API relationship, not sold as a monthly subscription.
- Yes. As of the pricing verified 2026-09-21, DeepSeek bills both deepseek-flash and deepseek-v4-pro at 50% of the peak rate outside its stated peak window of 01:00 to 04:00 and 06:00 to 10:00 UTC, Monday to Friday, a schedule that took effect 2026-09-10. The window has moved more than once, so check the DeepSeek pricing page before building batch schedules around it.
- Only above a fairly high sustained-throughput threshold. An 8x H100 node capable of serving the full DeepSeek V3 or R1 open-weight model rents for roughly $16-30 per hour, or approximately $12,000-22,000 per month running continuously. That only beats API spend once you are sustaining hundreds of millions to billions of tokens per month. For most teams the DeepSeek API is cheaper and involves less operational work.
- DeepSeek Chat throttled your account under the free tier's unpublished rate limit, the same limit that can degrade responses during heavy load. The error clears on its own once that throttle window passes. If it interrupts a task you cannot pause, move the workload to the DeepSeek API instead: API calls bill per token with no request-frequency cap, so a script or job that keeps tripping the free chat limit runs without interruption on the API.
- There is no paid upgrade for DeepSeek Chat's free-tier rate limit; DeepSeek does not sell a higher-quota consumer plan. The only paid path is the DeepSeek API, billed per token instead of by seat or quota. Create a key on the DeepSeek Platform, add billing, and send requests to the API instead of chat.deepseek.com. That removes the free chat rate limit entirely, and the per-token rates above show what it costs at your volume.
- No, DeepSeek publishes no official desktop app. DeepSeek Chat runs in a web browser at chat.deepseek.com, alongside official iOS and Android apps for mobile. There is no native Windows or macOS client, so a browser tab is the closest thing to a desktop client DeepSeek offers.
The complete AI playbook for your team
Cut your AI bill with Chinese open-weight models — without the risk: Safety, pricing and savings for Kimi K3, DeepSeek, Qwen and z.ai GLM — the four-vendor comparison for owners and IT leads.
Get the guide — $59 (reg. $89)