Deepseek has made its 75 percent pricing discount permanent on its V4-Pro model, undercutting OpenAI's GPT-4.5 by at least 34 times on output tokens. The move signals intensifying price competition in the AI market.
Deepseek is cementing aggressive pricing for its flagship V4-Pro model, making a temporary discount permanent. Input tokens now cost $0.435 per million—11.5 times cheaper than OpenAI's GPT-4.5. More significantly, output token pricing is at least 34 times lower than GPT-4.5's rates.
The pricing structure targets a critical market segment: agentic AI systems that consume massive token volumes. These systems generate their own outputs, compounding token usage. Lower output costs directly reduce operational expenses for developers building autonomous agents.
Deepseek's approach reflects a broader strategy to compete on cost and capability simultaneously. The Chinese AI company has positioned itself as a high-performance alternative to Western providers, with recent versions demonstrating competitive reasoning abilities at a fraction of the price.
For enterprise customers and startups evaluating AI infrastructure, the pricing differential is substantial. A system processing millions of tokens daily sees costs shift from prohibitive to manageable at these rates. This particularly impacts companies developing multi-step reasoning systems or extended interactions.
The permanent discount removes pricing uncertainty that typically accompanies temporary promotions. Customers can now build cost models around fixed rates rather than anticipating price increases.
Western AI providers face mounting pressure from this pricing model. OpenAI, Anthropic, and others have maintained relatively premium pricing despite competition. Deepseek's strategy forces a reckoning: either match prices and compress margins, or cede cost-sensitive segments to competitors.
The sustainability of Deepseek's pricing remains a variable. Lower prices require either superior operational efficiency, different margin expectations, or subsidization. The company's willingness to absorb lower per-token revenue suggests confidence in volume and market expansion.
For the AI infrastructure market, the competition now extends beyond model performance to economics. Agentic systems—the next frontier for AI deployment—make per-token costs a central decision factor. Deepseek's permanent pricing move accelerates the shift toward cost-driven adoption decisions.
Uber's weekly AI agent requests have grown nearly tenfold since February, yet the company has held spending flat since April after exhausting its entire 2026 AI budget in Q1.
A recent paper shows artificial intelligence often diagnoses and treats patients better than human physicians. The findings are prompting difficult conversations within the medical community about the profession's evolving role.
The Relay Q, launching next year, represents the latest push to establish voice as the primary interface for human-computer interaction, challenging the keyboard's decades-long dominance.
An Anthropic researcher demonstrated automated systems that can identify and correct misaligned behaviors without compromising overall performance. The systems improved on all 10 tested benchmarks measuring specific problematic outputs.