GPT-5 Pricing: OpenAI's Low Cost Could Trigger AI Price War

OpenAI Launches GPT-5 and Sparks Potential Price War

The technology sector was once again surprised by OpenAI this week with the unveiling of its latest flagship model, GPT-5. This launch occurred shortly after the company released a pair of new models under an open-source license.

CEO Altman's Assessment and Competitive Landscape

Sam Altman, CEO of OpenAI, confidently declared GPT-5 to be “the best model in the world.” However, assessments from sources like TechCrunch’s Maxwell Zeff suggest a more nuanced reality. GPT-5 demonstrates only marginal improvements over leading AI models developed by Anthropic, Google DeepMind, and xAI on certain benchmarks, and exhibits slight underperformance on others.

Despite this, the model delivers strong performance across a diverse range of applications, notably in the realm of coding. Altman also highlighted a key competitive advantage: pricing. He expressed satisfaction with the affordability of the new model via a recent social media post.

GPT-5 Pricing Structure

The GPT-5 API is priced at $1.25 per 1 million tokens for input and $10 per 1 million tokens for output. An additional $0.125 per 1 million tokens is charged for cached input. This pricing strategy closely aligns with Google’s Gemini 2.5 Pro basic subscription, which is also favored for coding tasks.

However, Google’s pricing structure becomes more expensive when input/output volumes exceed 200,000 prompts, potentially increasing costs for its highest-volume users.

Undercutting the Competition

OpenAI’s pricing significantly undercuts Anthropic’s Claude Opus 4.1, which begins at $15 per 1 million input tokens and $75 per 1 million output tokens. Anthropic does offer discounts for prompt caching and batch processing, which can mitigate costs.

Anthropic’s model is highly regarded among programmers, serving as a core component of coding assistants like Cursor and powering its own assistant, Claude Code. GPT-5 was made available as an option within Cursor almost immediately following its announcement.

Developer Feedback on Pricing

Early access developers have praised GPT-5’s pricing. Simon Willison, featured in OpenAI’s launch video, described the pricing as “aggressively competitive with other providers” in his review.

Matt Shumer, co-founder and CEO of OthersideAI (creator of HyperWrite), noted that GPT-5 “is cheaper than GPT-4o, which is fantastic. Intelligence per dollar continues to increase.”

Potential for a Price War

Reactions on platforms like X and Hacker News have been largely positive regarding OpenAI’s pricing. The question now is whether competitors like Anthropic and Google will respond with more affordable options.

A price war would be a welcome development. The economic viability of many AI-powered tool providers is currently precarious due to the high and unpredictable costs associated with model access, as reported by TechCrunch’s Marina Temkin.

Infrastructure Costs and Investment

The industry has been anticipating improvements in the LLM price-to-performance ratio and reductions in inference costs. However, such improvements seemed distant given the massive investments being made in data centers and infrastructure to support growing AI demand.

OpenAI has a substantial $30 billion annual contract with Oracle for capacity, while its annual recurring revenue recently reached $10 billion. Meta plans to invest up to $72 billion in AI infrastructure in 2025, and Alphabet has allocated $85 billion for capital expenditures in 2025, largely driven by AI needs.

A Gauntlet Thrown Down

Considering these significant investments, it may be premature for startups to celebrate OpenAI’s pricing adjustments. However, OpenAI has demonstrably challenged the existing price structure with not one, but two moves this week.

Whether other companies will respond remains to be seen.