DeepSeek V4 Pro Uncensored pricing, in full

The arithmetic for one request

Billing is per token, so a single request costs the sum of its three token classes. Here is a worked example at 1,000 uncached prompt tokens, 4,000 cached prompt tokens and 500 completion tokens — a typical follow-up turn in a long conversation.

Rounded to eight decimal places. The balance is held before the request and reconciled against the real usage afterwards, so the charge matches the tokens actually consumed.

What a month costs

Four modelled traffic shapes at this rate card. These are representative token shapes, not benchmarks — substitute your own volumes on the calculator.

Uncensored vs standard, same price

The uncensored build is not a premium tier here. It is priced identically to the standard V4 Pro, so the choice is about behaviour rather than budget.

Every refusal-free option, priced

The uncensored and abliterated models in this catalog, sorted cheapest input first. This is the table to read if the question is which no-refusal model to run rather than which DeepSeek tier.

Within the DeepSeek family

All 8 DeepSeek models on this gateway, cheapest input first. The premium for the pro tier over the cheapest model is 12.5× on output.

Send a request

One endpoint, one key. The same request shape works for every model in the catalog.

curl https://uncensoredai.dev/v1/chat/completions \
  -H "Authorization: Bearer $UNCENSOREDAI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-pro-uncensored",
    "messages": [{ "role": "user", "content": "Hello" }],
    "stream": true
  }'

Questions

How much does DeepSeek V4 Pro Uncensored cost?

$0.50 per 1M input tokens, $0.10 per 1M cached input tokens and $0.50 per 1M output tokens, billed per token from a prepaid balance. There is no subscription, seat fee or minimum spend.

How does it compare with the standard DeepSeek V4 Pro?

Both are priced identically on this gateway. The difference is the weights, not the rate card: the standard build keeps the provider's refusals, while the uncensored build has them relaxed at the provider tier.

What is the difference between uncensored and abliterated?

Uncensored means the provider relaxed the safety refusals around the standard weights. Abliterated means the refusal direction was removed from the weights themselves, so the change holds across every request rather than depending on how the model was deployed.

How much does a million tokens actually cost?

At this rate card, 1M input tokens costs $0.50 and 1M output tokens costs $0.50. A request with 1,000 input and 500 output tokens therefore costs about $0.000750.

Is there a free tier?

No. There is no free credit and no trial balance; the account starts at zero and is funded by a crypto deposit. Generating a key is free and instant, but a request needs a positive balance.

Related

Rates verified 2026-10-26 · cheapest model here is DeepSeek V4.1 Flash at $0.04/1M input