Free DeepSeek API

Free DeepSeek API. Exact model ID. Live availability.

DeepSeek V4.1 Flash Economy through an OpenAI-compatible endpoint. The model is free while the published window is open. This page shows the live gate, the exact model ID, the free limits and copy-paste setup for the tools you already use.

OpenAI-compatible 1M context $0.50 signup credit No card at signup Cline, Roo Code, Open WebUI

Free window closed Checked 2026-11-18T14:21:12.998742+00:00. Availability is controlled by the live gate, not a promise.

Model ID

deepseek-v4-1-flash-economy-free

Paste this exact string into any OpenAI-compatible client.

Price during the window

$0.00

Per 1M tokens. Requests are not charged to your wallet while the gate is open.

Window ends

2026-11-22

30 billion tokens are up for grabs until November 22, provided the DeepSeek cache hit rate stays at 95% or better. Below that, the giveaway ends sooner. Whichever limit trips first.

Live availability

Is the free model open right now?

This is read from the same server-side gate that routes the model. It is not a static marketing claim.

Status

Closed

The free alias is not accepting new requests. Paid model IDs still work.

Tokens up for grabs

30 billion tokens are up for grabs until November 22, provided the DeepSeek cache hit rate stays at 95% or better.

2.6B of 30B tokens used so far.

Use the paid model

Free usage limits

What you get without topping up.

The free model has no token-per-minute or token-per-day cap. Other models follow the standard limits. A top-up is not required to use it.

LimitFree accountAfter any top-up
Requests per minute (free model only)1,0003,000
Requests per day (free model only)100,000250,000
Tokens per minuteNo limitNo limit
Tokens per dayNo limitNo limit
Requests per minute (other models)303,000
Requests per day (other models)300250,000
Tokens per minute (other models)100,000No limit
Tokens per day (other models)1,000,000No limit
Free model token limitsNo TPM/TPD capNo TPM/TPD cap
Free model price$0.00 / 1M$0.00 / 1M

Limits come from the live account configuration. They are not estimates. The free window is shared until November 22 or 30B tokens, whichever comes first.

Quickstart

The exact model ID, in your language.

Base URL: https://api.fikraapi.co.ke/v1. API key: your fk_live_ key. Model: deepseek-v4-1-flash-economy-free.

Python

pip install openai
from openai import OpenAI

client = OpenAI(base_url="https://api.fikraapi.co.ke/v1", api_key="fk_...")

response = client.chat.completions.create(
    model="deepseek-v4-1-flash-economy-free",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

JavaScript

npm install openai
import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.fikraapi.co.ke/v1", apiKey: "fk_..." });

const response = await client.chat.completions.create({
  model: "deepseek-v4-1-flash-economy-free",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(response.choices[0].message.content);

cURL

curl https://api.fikraapi.co.ke/v1/chat/completions \
  -H "Authorization: Bearer fk_..." \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-v4-1-flash-economy-free","messages":[{"role":"user","content":"Hello"}]}'

Use the paid model

Tool setup

Use it in Cline, Roo Code and Open WebUI.

Every step below uses the tool's current OpenAI-compatible settings. No provider-specific adapter is needed.

Cline

  1. Install Cline in VS Code and open its settings panel with the gear icon.
  2. Set API Provider to OpenAI Compatible.
  3. Set Base URL to https://api.fikraapi.co.ke/v1.
  4. Set API Key to your fk_live_ key.
  5. Set Model ID to deepseek-v4-1-flash-economy-free.
  6. Save and start a task.

Roo Code

  1. Install Roo Code in VS Code and open its settings panel.
  2. Set API Provider to OpenAI Compatible.
  3. Set Base URL to https://api.fikraapi.co.ke/v1.
  4. Set API Key to your fk_live_ key.
  5. Set Model ID to deepseek-v4-1-flash-economy-free.
  6. Save and start a task.

Open WebUI

  1. Open Settings > Admin > Connections.
  2. Under Manage OpenAI API Connections, click Add Connection.
  3. Set URL to https://api.fikraapi.co.ke/v1 and the API Key to your fk_live_ key.
  4. Click Verify Connection. If model auto-detection fails, add deepseek-v4-1-flash-economy-free to the Model IDs allowlist.
  5. Select the model in a new chat.

When the free window ends.

The free model is a time-limited promotional window, not a permanent free tier. There are two close conditions, and the first one to happen wins.

  • Time: 2026-11-22.
  • Tokens: 30 billion tokens are up for grabs until November 22, provided the DeepSeek cache hit rate stays at 95% or better.
  • Fail closed: when either condition trips, the alias stops accepting new requests. There is no auto-renewal.

Nothing already accepted during the window is charged to your wallet. After it closes, the same account and base URL work with any paid model ID.

Keep going after the window

Top up from $5.00. Top-ups are charged in KES at KES 133 = $1 with a 3% fee ($5 minimum; KES 136.99 charged, KES 133.00 credited). Your wallet is credited in USD, and paid limits activate on the first top-up.

Existing paid model IDs keep the same endpoint and the same key. No migration is needed.

Use the paid model See paid prices

FAQ

Free DeepSeek API questions.

Is it really free?

While the live gate is open, requests to deepseek-v4-1-flash-economy-free are not charged to your wallet. The window is shared and time limited.

Do I need a card?

No. Signup takes one step and starts with $0.50 in credit. A card is only needed when you choose to top up.

What is the exact model ID?

deepseek-v4-1-flash-economy-free. It is case sensitive.

Which tools work?

Any OpenAI-compatible client, including Cline, Roo Code, Open WebUI, Continue, Aider and the official OpenAI SDKs.

What are the limits?

The free model gets 1,000 requests/minute and 100,000 requests/day on the free model, with no token cap (other models: 30 requests/minute, 300 requests/day). Any top-up activates paid limits: 3,000 requests/minute and 250,000 requests/day, no token ceilings, on every model.

What happens after expiry?

The free alias stops accepting requests. Your account, key and base URL stay the same; switch the model ID to a paid DeepSeek build to keep going.

Related: DeepSeek line · Live pricing · Model catalog · Free window announcement · Public stats · /status · /support. Operated by Lacesse Ventures. No contractual SLA on the self-serve plan.

Get the key, run the first call.

Create an account, copy the snippet, and send your first request to the free model. The exact model ID is deepseek-v4-1-flash-economy-free.

Use the paid model