DeepSeek Makes 75% Discount on V4 Pro Permanent — Prices Fall 4x

DeepSeek made the 75% discount on V4 Pro permanent. The price fell from $0.0145 to $0.003625 per million input tokens — eight times cheaper than GPT-5. The promotion was temporary until May 31, but the company is keeping it forever. This is a serious blow to industry pricing.

AI-processed from TNW; edited by Hamidun News
DeepSeek Makes 75% Discount on V4 Pro Permanent — Prices Fall 4x
Source: TNW. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTNW ↗
Editor-in-Chief
Key metric #1V4 Pro took top positions in global benchmarks
Key metric #2DeepSeek became 4-8 times cheaper than market leaders
Key metric #3Companies started reviewing their contracts for OpenAI model usage

DeepSeek has announced an extension of the 75% discount on its flagship V4 Pro model indefinitely. Last week, this offer was supposed to end on May 31, but the company decided to make it permanent.

Numbers That Shock

The new DeepSeek prices look almost incredible. Input tokens now cost $0.003625 per million, output tokens — $0.087. Previously, they were four times more expensive: $0.0145 and $0.35 respectively. For comparison, OpenAI's GPT-5 costs $2.50 per million input tokens — 689 times more expensive. Even OpenAI and Anthropic's budget models remain noticeably more expensive than DeepSeek. Claude 3.5 Sonnet costs $3 per input tokens. Grok is in a similar price range to OpenAI.

The Price War Reached the Core

DeepSeek's decision to keep the discount permanent is not just a PR move. It's a signal that the Chinese startup intends to stay in the game as aggressively cheap as all competitors. The offer was originally announced as a limited-time promotion, but judging by the extension, competition is forcing the company to take this step. Here's what happened:

  • V4 Pro took top positions in global benchmarks
  • DeepSeek became 4-8 times cheaper than market leaders
  • Companies started reviewing their contracts for OpenAI model usage
  • Competitors are forced to lower prices or lose customers

A New Phase of Competition

The world of AI models has entered a new phase: this is no longer a war of features, but a war of prices. DeepSeek shows that quality models can be made cheaply — and profitably. Large companies that are currently overpaying for GPT-4 or GPT-5 will start counting every penny.

What This Means

AI prices will fall even further. If OpenAI, Anthropic, and other leaders do not respond aggressively to this move, they will lose some clients by autumn. At the same time, DeepSeek's low cost may become a barrier to its adoption in the corporate sector — companies usually choose proven solutions, not just based on price.

Story

Key Questions Answered

What are the key details on Numbers That Shock?

The new DeepSeek prices look almost incredible. Input tokens now cost $0.003625 per million, output tokens — $0.087. Previously, they were four times more expensive: $0.0145 and $0.35 respectively. For comparison, OpenAI's GPT-5 costs $2.50 per million input tokens — 689 times more expensive. Even OpenAI and Anthropic's budget models remain noticeably more expensive than DeepSeek. Claude 3.5 Sonnet costs $3 per input tokens. Grok is in a similar price range to OpenAI.

What are the key details on The Price War Reached the Core?

DeepSeek's decision to keep the discount permanent is not just a PR move. It's a signal that the Chinese startup intends to stay in the game as aggressively cheap as all competitors. The offer was originally announced as a limited-time promotion, but judging by the extension, competition is forcing the company to take this step. Here's what happened: V4 Pro took top positions in global benchmarks DeepSeek became 4-8 times cheaper than market leaders Companies started reviewing their contracts for OpenAI model usage Competitors are forced to lower prices or lose customers

What are the key details on A New Phase of Competition?

The world of AI models has entered a new phase: this is no longer a war of features, but a war of prices. DeepSeek shows that quality models can be made cheaply — and profitably. Large companies that are currently overpaying for GPT-4 or GPT-5 will start counting every penny.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…