Price checkPer million tokens
Free ChatFreeAPI Access (V4 Flash)CustomAPI Access (V4 Pro)Custom
See all 3 plans

All DeepSeek Plans & Pricing

Prices unverified — last checked Jul 27, 2026

Plan Monthly Annual Best For
View all features by plan (compare side-by-side)

Free Chat

  • Unlimited chat messages
  • File uploads
  • Daily reset quotas
  • V4 model access

API Access (V4 Flash)

  • deepseek-flash (DeepSeek-V4.1-Flash): non-thinking and thinking modes
  • Input tokens: $0.30/1M peak, $0.15/1M off-peak (cache miss)
  • Cache hit input: $0.006/1M peak, $0.003/1M off-peak
  • Output tokens: $1.20/1M peak, $0.60/1M off-peak
  • Off-peak = all hours except Mon–Fri 01:00–04:00 and 06:00–10:00 UTC
  • 1M context length
  • 384K max output tokens
  • JSON output, tool calls, vision, FIM completion (non-thinking)

API Access (V4 Pro)

  • deepseek-v4-pro (DeepSeek-V4-Pro-0813): non-thinking and thinking modes
  • Input tokens: $1.32/1M peak, $0.66/1M off-peak (cache miss)
  • Cache hit input: $0.044/1M peak, $0.022/1M off-peak
  • Output tokens: $3.96/1M peak, $1.98/1M off-peak
  • 1M context length
  • 384K max output tokens
  • JSON output, tool calls, FIM completion (non-thinking)
Pricing Alerts

Track DeepSeek pricing

Get an email when DeepSeek's pricing changes — plus the weekly SaaS Price Watch: verified price changes and deals across 3,000+ products. One-click unsubscribe.

Try DeepSeek Free
Compare DeepSeek with alternativesAdjust seats, lock a tier, add up to 2 more products side-by-side. Shareable URL.
Quick Answer
Last verified:
High confidence

DeepSeek costs Free to Free per million tokens as of November 2026, with 3 plans available including a free tier. Plan: Free Chat (free). Custom pricing is available on request. Pricing depends on your chosen tier, contract length, and negotiated discounts.

Use the interactive pricing calculator to estimate your exact cost based on team size and requirements.

  • Free tier: Yes

DeepSeek offers 3 pricing tiers: Free Chat, API Access (V4 Flash), API Access (V4 Pro). The API Access (V4 Flash) plan is developers needing fast, low-cost api integration.

Compared to other ai chatbots software, DeepSeek is positioned at the budget-friendly price point.

  • 4 documented hidden costs beyond list price
Last verified: | Sources: 7 pricing pages and reviews

How much does DeepSeek cost?

DeepSeek uses usage-based pricing from $0.006 per 1M tokens, with a free tier. Enterprise pricing is quote-based. Plans include Free Chat (free), API Access (V4 Flash) (custom pricing), API Access (V4 Pro) (custom pricing).

DeepSeek Pricing Overview

DeepSeek has 3 pricing plans, including a free tier. Paid plans range from $0 to $0/per million tokens. The Free Chat plan is free and is best for personal use and experimentation. The API Access (V4 Flash) plan requires contacting sales for a custom quote and is designed for developers needing fast, low-cost api integration. The API Access (V4 Pro) plan requires contacting sales for a custom quote and is designed for production workloads needing higher-capability model.

DeepSeek with a None — prepaid balance model, add funds as needed minimum commitment.

There are at least 4 documented hidden costs beyond DeepSeek's list price, including implementation, training, and add-on fees.

This pricing was last verified in July 27, 2026 from 7 independent sources.

Try DeepSeek Free

DeepSeek offers free chat access and pay-per-token API pricing starting at $0.15 per million input tokens (off-peak) as of October 2026. The Free Chat tier provides access to DeepSeek models with no subscription required. The API's deepseek-flash model costs $0.30 per million input tokens and $1.20 per million output tokens at peak rates, and half that off-peak; deepseek-v4-pro costs $1.32/$3.96 at peak. Verified from DeepSeek's API pricing page by Costbench, the software pricing database tracking 1,000+ products.

How DeepSeek Pricing Compares

Compare DeepSeek pricing against top alternatives in AI Chatbots.

Character.ai $0-$119.88/month Compare → ChatGPT $0-$100/user/month Compare → Chatsonic $19-$299/month Compare →

Usage-Based Rates

Per-unit pricing for DeepSeek API usage.

API Access (V4 Flash)

Model Input Output Cached Per
deepseek-flash 1000K ctx $0.300 $1.20 $0.006 1M tokens
  • Rates shown are peak rates; off-peak rates (all hours except Mon–Fri 01:00–04:00 and 06:00–10:00 UTC) are half
  • Legacy model names deepseek-v4-flash and deepseek-v4-flash-vision-exp are served by DeepSeek-V4.1-Flash-1017 and billed at the Flash price
  • Supports both thinking (default) and non-thinking modes

API Access (V4 Pro)

Model Input Output Cached Per
deepseek-v4-pro 1000K ctx $1.32 $3.96 $0.044 1M tokens
  • Rates shown are peak rates; off-peak rates (all hours except Mon–Fri 01:00–04:00 and 06:00–10:00 UTC) are half
  • Supports both thinking (default) and non-thinking modes

Compare DeepSeek vs Alternatives

Before committing to DeepSeek, compare pricing with these 3 alternatives in the same category.

VSCharacter.ai
Free

Casual users and those exploring AI chatbots

Full comparison
VSChatGPT
Free

Casual users exploring ChatGPT with light usage needs

Full comparison
VSChatsonic
From $19/month

Small teams and individuals

Full comparison

All DeepSeek alternatives & migration guides

What Companies Actually Pay for DeepSeek

Median per-1M-token pricing across 11 models
Input $0.290/1M
Output $0.790/1M
Flagship models in this provider's catalog
Model Input /1M Output /1M Blended /1M
deepseek/deepseek-r1 $0.700 $2.50 $1.60
deepseek/deepseek-r1-0528 $0.500 $2.15 $1.32
deepseek/deepseek-v3.2-speciale $0.400 $1.20 $0.800
deepseek/deepseek-r1-distill-llama-70b $0.700 $0.800 $0.750
deepseek/deepseek-chat $0.320 $0.890 $0.605
Review scores
Third-party review aggregates, as of Apr 2026
Top pricing complaints
Rate limiting and single-request-at-a-time queuing on direct API during peak periodsPeriodic service disruptions and paused fund additions during high-demand periodsContext window limited to 64K tokens on direct API compared to larger windows via third-party providersData privacy concerns related to Chinese server infrastructure
Source: OpenRouter API — medians aggregated from 11 models routed. Reflects router-surface pricing (may include modest markup vs direct provider rates).

DeepSeek Cost by Company Size

Example costs for teams of different sizes. Each figure is labelled with the period it covers.

Personal Chat Usage Free

Individual users exploring AI capabilities with unlimited free access

Small Developer Project $0.15-1.20 per month

$0.15-1.20/month (deepseek-flash, off-peak input to peak output rates)

Processing ~1M tokens monthly for a small application or prototype

Medium Business Integration $1.50-12 per month

$1.50-12/month (deepseek-flash, off-peak input to peak output rates)

Processing ~10M tokens monthly for customer service or content generation

Enterprise API Usage $15 estimated cost

$15+ per month (100M input tokens at the $0.15/1M off-peak rate; cache hits cost less)

High-volume processing with 100M+ tokens monthly using cache optimization

Solo Developer — Light API Usage $2 estimated cost

Under $2 for 15M tokens

A single developer using DeepSeek API for coding assistance, dropping approximately 15M total tokens

Per-Project Coding Use — Pay-As-You-Go $2-$5 estimated cost

$2-$5 per project

Typical cost per individual coding or QR-code generation project using the API pay-as-you-go, across various models including DeepSeek R1

How DeepSeek Pricing Compares

Software Starting Price Top Price
DeepSeek Custom Custom
Character.ai Free $7.92/month
ChatGPT Free $100/user/month
Chatsonic $19/month $299/month
Claude Free $200/user/month
Cohere Free $4300/month

Detailed pricing comparisons:

Browse all AI Chatbots pricing →

4 DeepSeek Hidden Costs Beyond the List Price

Beyond the listed price, DeepSeek has at least 4 documented hidden costs that can significantly increase total cost of ownership.

Watch for 4 hidden costs
  • Chain-of-Thought Reasoning Token Costs (Thinking Mode) 50-70% of total thinking-mode API costs for complex reasoning tasks
    high 1 source
    Reddit “Every token used in the CoT reasoning and the final answer is charged at $2.19 per million tokens, as it is part of the output. It is a bit of a trade-off; more CoT tokens = more cost but also better accuracy in complex tasks.”
    Reddit “We reduced costs by about 57% compared to the original with the optimized prompt, while still getting the correct answer.”
  • Cache Miss Penalty — 30-50x Higher Input Costs Varies with cache hit rate — cached input costs 2-3% of the cache-miss price
    medium 2 sources
    Reddit “STANDARD PRICE (UTC 00:30-16:30) 1M TOKENS INPUT (CACHE HIT) $0.07 | 1M TOKENS INPUT (CACHE MISS) $0.27 | 1M TOKENS OUTPUT $1.10”
    Reddit “caching saves us about 74.07% in API costs for deepseek-chat and 74.55% for deepseek-reasoner when compared to the cost of processing new (uncached) input tokens.”
  • Massive Provider Markup on Third-Party Platforms 10-100x vs cheapest available provider for same model
    high 1 source
    Hacker News “Take DeepSeek R1 0528 (quality 68 from Artificial analysis bench, beats many flagships): Completely free on Google Vertex and CentML (decent speeds too, 121 tok/s and 87 tok/s). But jumps to $0.91 on Deepinfra, $4.”
  • Service Availability During High Demand 5-15% of license costs
    medium 1 source
    Reddit “At the time of publishing this article, DeepSeek has paused adding new funds to API accounts because their servers are running low on resources. That means that you can't add more money to your API balance right now.”
Tip

Ask your DeepSeek sales rep about these costs upfront. Getting them in writing before signing can save you from surprise charges later.

Full hidden costs breakdown →

Intelligence sourced from 2 independent sources
Reddit User discussions Hacker News Tech community
Key claims include inline source attribution. Data verified against multiple independent sources. 16 source citations total.

DeepSeek Contract Terms

DeepSeek contracts do not auto-renew. Changes require advance notice. These terms are sourced from verified buyer experiences.

Contract Terms
Auto-Renewal No
Minimum Commitment None — prepaid balance model, add funds as needed
Mid-Term Downgrade Allowed
Payment Terms Prepaid balance consumed per-token; no invoicing or net terms
Price Escalation No published schedule; historically pricing has decreased as newer model versions are released
Note

No tiers to downgrade between; simply switch to a cheaper model (deepseek-flash vs deepseek-v4-pro) or reduce usage

Based on 1 verified source

How to Negotiate DeepSeek Pricing

DeepSeek contracts are negotiable. These 8 tactics are sourced from real buyer experiences and procurement specialists.

Negotiation Playbook 8 tactics
API Pricing is Already Among the Lowest in the Market high success

DeepSeek API list pricing is 80-95% cheaper than GPT-4o and Claude equivalents. At these rates, negotiating further has limited ROI — focus instead on controlling input/output token ratios (shorter prompts, cached context) to cut costs by 30-50% without any vendor negotiation.

DeepSeek vs competitor API pricing
Cache Repeated Prompts to Minimize Token Cost high success

DeepSeek supports prompt caching for repeated system prompts and context. Enabling caching on shared system prompts (e.g., same instructions sent with every request) can cut the cost of cached input tokens by about 97-98% (for example $0.006 vs $0.30 per 1M on deepseek-flash at peak).

DeepSeek API documentation
Schedule Batch Jobs During Off-Peak Hours high success

DeepSeek's direct API automatically charges half the peak rate during off-peak hours — all hours except Mon–Fri 01:00–04:00 and 06:00–10:00 UTC. This requires no negotiation — simply schedule non-urgent batch processing, data analysis, or document work outside the peak windows. deepseek-v4-pro output drops from $3.96 to $1.98/1M tokens off-peak.

DeepSeek API pricing page (October 2026)
Maximize Context Cache Hit Rate high success

Cached input tokens cost about 98% less than uncached inputs ($0.006 vs $0.30/1M for deepseek-flash at peak). Keep static content — system prompts, reference documents, few-shot examples — at the beginning of every request. The system automatically caches inputs of 64+ tokens and handles it transparently; no code changes are needed beyond prompt structure.

Reddit (SurferCloudServer API guide, April 2025)
Use deepseek-flash and Non-Thinking Mode for Routine Tasks high success

deepseek-v4-pro costs about 3-4x more per token than deepseek-flash ($1.32 vs $0.30 input and $3.96 vs $1.20 output per 1M at peak), and thinking mode adds chain-of-thought output tokens. For code generation, content writing, data extraction, and general chat, deepseek-flash in non-thinking mode delivers strong results at a fraction of the price. Reserve V4 Pro or thinking mode for tasks that need multi-step reasoning.

DeepSeek API pricing page (October 2026)
Constrain Output Format in Prompts high success

Adding explicit output constraints to prompts (e.g., 'Show formula and result only' instead of an open-ended question) reduces output token usage by over 50% in documented examples. In thinking mode, this also shortens chain-of-thought, cutting the most expensive token type. Structure every prompt to request the minimum viable output.

Reddit (SurferCloudServer API guide, April 2025)
Compare Providers Before Committing to a Platform high success

The same DeepSeek model can cost anywhere from free (Google Vertex, CentML during their free tiers) to $5.50/1M on SambaNova. Before integrating into production, benchmark at least three providers on OpenRouter or directly. For experimentation and batch runs, free provider tiers can entirely eliminate inference costs.

Hacker News (demian101, July 2025)
Use Direct DeepSeek API vs OpenRouter for Lowest Base Prices high success

Direct API pricing is consistently lower than third-party router markups, especially combined with off-peak discounts. Community consensus is that 'no provider can compete in pricing against DeepSeek's off peak hours' on the direct API. For sustained production usage, direct API access provides the best cost floor.

Reddit (DeepSeek subreddit, May 2025)
High success Medium Low / difficult

Full negotiation guide →

DeepSeek Pricing FAQ

01 How much does DeepSeek cost?

DeepSeek offers free chat access. API pricing starts at $0.15 per million input tokens (deepseek-flash, off-peak, cache miss; $0.30 at peak), with output from $0.60 per million tokens off-peak ($1.20 at peak).

02 Does DeepSeek have a free plan?

Yes, DeepSeek provides completely free chat access with unlimited messages and file uploads within daily quotas. No subscription required.

03 What is DeepSeek's API pricing?

deepseek-flash costs $0.30 input / $1.20 output per million tokens at peak and half that off-peak; deepseek-v4-pro costs $1.32 / $3.96 at peak ($0.66 / $1.98 off-peak). Cache hits cost $0.006 (Flash) and $0.044 (V4 Pro) per million input tokens at peak.

04 Are there any subscription plans?

No, DeepSeek doesn't offer traditional subscription plans like 'Plus' or 'Pro'. It's either free chat or pay-per-use API access.

05 How does DeepSeek compare to competitors pricing?

DeepSeek is among the most affordable AI APIs available, significantly cheaper than OpenAI GPT and Claude models while offering competitive performance.

06 What are the free token limits?

DeepSeek's API pricing page refers to a granted balance that is used before your topped-up balance, but it does not publish a standard free-token amount for new API accounts. Chat users get daily quotas that reset each day.

07 Is DeepSeek free to use?

The Free Chat plan provides unlimited access to the DeepSeek chat interface at no cost. API Access is pay-per-token with no monthly fee — you prepay a balance and consume it as you make API calls. There is no free tier for the API itself, but actual costs are very low: 15 million tokens costs under $2 on the direct API.

08 What is the cost difference between DeepSeek's Flash and Pro API models?

DeepSeek no longer prices V3 (deepseek-chat) and R1 (deepseek-reasoner) separately. Its current API models are deepseek-flash at $0.30 input / $1.20 output per million tokens and deepseek-v4-pro at $1.32 / $3.96 (peak rates; off-peak is half), so V4 Pro costs roughly 3-4x more per token. Both support a thinking mode whose chain-of-thought tokens are billed as output, so reasoning-heavy queries cost more than the visible answer suggests.

09 Does DeepSeek offer off-peak pricing discounts?

Yes. Off-peak rates are half of peak rates on both deepseek-flash and deepseek-v4-pro. Peak hours are 01:00–04:00 and 06:00–10:00 UTC, Monday through Friday (excluding Chinese public holidays); all other hours, including weekends, are off-peak. For example, deepseek-v4-pro output drops from $3.96 to $1.98 per million tokens off-peak. The discount applies automatically on the direct API — no configuration needed.

10 How does DeepSeek context caching work and how much can it save?

DeepSeek automatically caches input text chunks of 64 tokens or more. Cached inputs cost about 98% less than uncached inputs on deepseek-flash ($0.006 vs $0.30/1M at peak) and about 97% less on deepseek-v4-pro ($0.044 vs $1.32/1M). Cache usage is tracked in API responses via `prompt_cache_hit_tokens` and `prompt_cache_miss_tokens` fields. Cache expires after a few hours to days if unused. To maximize savings, keep system prompts and reference documents at the start of every request.

Is this pricing incorrect? — we'll verify and update it.