You want to try DeepSeek V4 — the model family that reset open-weights pricing on April 24, 2026 — but you are not sure which “DeepSeek” site is the real one, whether the chatbot and the developer console use the same login, or what the free tier actually covers. A clean DeepSeek account setup answers all three questions in about ten minutes, and this guide walks through every step on web, mobile, and the API platform.

You will get the verified URLs, the exact sign-in options, the difference between the chat account and the developer console, and a short checklist for keeping the account secure. The end of the article includes a worked cost example so you know what billing looks like before you commit a payment method.

The short answer: two separate accounts, one company

DeepSeek runs two distinct sign-in surfaces, and getting them confused is the most common setup mistake. The chat account at chat.deepseek.com gets you into the consumer chatbot and the mobile app. The developer account at platform.deepseek.com gets you into the API console where you create API keys, top up a balance, and read usage logs. They share branding and a parent company, but the credentials are not interchangeable — you sign up for each one separately.

If you only want to chat with the model, you need the first account. If you plan to call the API from code or build a product, you need the second one (and you can do both with the same email, just registered twice).

Verified entry points (April 2026)

• Web chat: chat.deepseek.com — sign-up page is chat.deepseek.com/sign_up, which describes itself as the chat interface for coding, content creation, file reading, and long-context conversations.
• Developer platform: platform.deepseek.com/sign_up — the DeepSeek API platform for accessing AI models, developer resources, and API documentation.
• Mobile app: install only from the Apple App Store, Google Play, or links on deepseek.com. Sideloaded APKs that claim to be DeepSeek are a known phishing vector.

Anything else — deepseek.ai, deep-seek.ai, chat-deep.ai — is a third-party site, not DeepSeek. Some are useful independent guides; none of them can manage your real account, password, or API keys. If you have any doubt, our verify official DeepSeek app guide lists every legitimate domain.

Step 1 — Create the chat account

1. Open chat.deepseek.com/sign_up in a current browser (Chrome, Firefox, Edge, or Safari).
2. Choose a sign-in method. Email and Google are available globally; phone and other methods may appear depending on region. Use the same one each time you log in.
3. Confirm the verification code if you registered with email or phone.
4. Accept the Terms and Privacy Policy. Signing up or logging in counts as consent to DeepSeek’s Terms of Use and Privacy Policy.
5. You are now in the chat interface. The default model is V4; toggle Expert Mode for V4-Pro-grade answers or Instant Mode for faster V4-Flash responses.

That’s the entire consumer flow. No payment details are required. For more on what the chatbot can do once you are in, see our walk-through on how to use DeepSeek.

What “V4” means in the chat UI

On April 24, 2026, DeepSeek released the V4 Preview as DeepSeek-V4-Pro at 1.6T total / 49B active parameters and DeepSeek-V4-Flash at 284B total / 13B active parameters. Both models are accessible at chat.deepseek.com via Expert Mode and Instant Mode. The DeepThink toggle still exists; on V4 it switches the model between non-thinking and thinking mode rather than swapping models. For a deeper look at the family, see DeepSeek V4, DeepSeek V4-Pro, and DeepSeek V4-Flash.

Step 2 — Create the developer (API) account

If you want to call the API from code, register a second account at platform.deepseek.com. The flow is similar:

1. Open platform.deepseek.com/sign_up.
2. Register with email or one of the supported social options.
3. Verify the address.
4. Set up two-factor authentication on the security page (recommended for any account that will hold an API key).
5. Open the API Keys tab, click Create new API key, copy the secret immediately, and store it in a password manager. The key is shown only once.

For a full annotated walkthrough including key naming conventions and rotation, see our get a DeepSeek API key page.

The smallest possible test call

Once you have a key, the simplest way to confirm everything works is a one-line model swap on the OpenAI Python SDK. Chat requests hit POST /chat/completions, the OpenAI-compatible endpoint at https://api.deepseek.com. DeepSeek also ships an Anthropic-compatible surface against the same base URL, so the Anthropic SDK works too.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.deepseek.com",
    api_key="sk-...",  # the key you just created
)

resp = client.chat.completions.create(
    model="deepseek-v4-flash",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(resp.choices[0].message.content)

If that prints a sentence, your account, billing, and key are all working. The API is stateless — every call must resend the full conversation history if you want multi-turn context, which is the main behaviour difference from the chatbot. DeepSeek’s own migration guidance is to keep base_url and just update model to deepseek-v4-pro or deepseek-v4-flash; both support OpenAI ChatCompletions and Anthropic APIs, with 1M context and dual modes (thinking / non-thinking).

Legacy model IDs

If you are migrating from older code, the legacy IDs deepseek-chat and deepseek-reasoner still work for now. deepseek-chat and deepseek-reasoner will be fully retired and inaccessible after July 24, 2026, 15:59 UTC; they currently route to deepseek-v4-flash in non-thinking and thinking modes respectively. Plan a one-line swap to deepseek-v4-flash or deepseek-v4-pro well before that date.

Step 3 — Add billing (developer account only)

The chatbot is free; the API is pay-as-you-go. Open the Billing tab on the developer console, add a card, and top up a starting balance ($5–$10 is plenty for testing). DeepSeek may show a granted balance — a small promotional credit that can expire — in the same view; check the billing console for current offers rather than assuming one is available.

What you will pay (as of April 25, 2026)

DeepSeek’s published pricing for the V4 tier is significantly lower than the previous V3.2 rates. DeepSeek’s pricing page lists $0.14 per million input tokens and $0.28 per million output tokens for Flash, and $1.74 per million input and $3.48 per million output for Pro. Cache-hit input is cheaper still. Always confirm against the live DeepSeek API pricing page before scaling.

Off-peak discounts are not active — DeepSeek discontinued them on September 5, 2025, and did not reintroduce them with V4.

Worked example: 1,000,000 calls on V4-Flash

Assume a 2,000-token system prompt cached across calls, a 200-token user message (uncached), and a 300-token response.

• Cached input: 2,000 × 1,000,000 = 2,000,000,000 tokens × $0.028/M = $56.00
• Uncached input: 200 × 1,000,000 = 200,000,000 tokens × $0.14/M = $28.00
• Output: 300 × 1,000,000 = 300,000,000 tokens × $0.28/M = $84.00
• Total: $168.00

The same workload on V4-Pro costs $290 + $348 + $1,044 = $1,682. For most chat and content workloads, Flash is the right default; reach for Pro only when frontier-tier coding or agentic benchmarks justify the spend.

Step 4 — Lock the account down

Before you do anything else, treat the developer account like a payment account, because it is one. A leaked API key can run up real charges before you notice.

• Enable two-factor authentication on both the chat and platform accounts.
• Use a password manager. Reusing a password between DeepSeek and a breached site is the most common way accounts get hijacked.
• Name your API keys by environment (e.g. local-dev, staging, prod) so you can revoke one without breaking the others.
• Never commit keys to Git. Use environment variables or a secret manager.
• Rotate keys on a schedule — quarterly is a sensible baseline for production keys.

If the worst happens and you lose access, our reset DeepSeek password guide covers the recovery flow on both surfaces.

Step 5 — Pick the right surface for your workflow

The web and app keep conversation history on the server for your session, so you can return to past chats. The API does not — your application is responsible for storing and replaying the message array. If you are deciding between surfaces, our DeepSeek browser vs app comparison goes into more depth.

Edge cases and country availability

Sign-up is not uniform worldwide. Several countries have banned government use of DeepSeek, including Italy, the United States, and South Korea, citing national security concerns; Germany also blocked DeepSeek in the Apple and Google app stores in 2025, citing illegal transfer of user data to China. Consumer use in these countries is generally still possible, but app-store availability and corporate policies vary. The full breakdown lives on our DeepSeek availability by country page.

Common edge cases:

• Email domain rejected: some disposable-email providers are blocked. Use a primary mailbox.
• Phone verification fails in your region: switch to email or Google sign-in.
• App not on your local store: use the web chat at chat.deepseek.com in a browser.
• Corporate firewall blocks api.deepseek.com: open a ticket with IT; do not route through unofficial proxies.

Related guides

Once your account works, the next reading depends on what you are building. The DeepSeek beginner guides hub indexes everything; for specific paths, our DeepSeek API getting started tutorial picks up where this guide ends, and the DeepSeek OpenAI SDK compatibility reference covers the wire-format details.

Last verified: 2026-04-25. DeepSeek AI Guide is an independent resource and is not affiliated with DeepSeek or its parent company. Model IDs, pricing and API behaviour change; check the official DeepSeek documentation and pricing page before committing to a production decision.