OpenAI priced GPT-5 so low, it may spark a price war
OpenAI has once again captured the tech industry’s attention, this time with the launch of its highly anticipated flagship model, GPT-5. This release follows closely on the heels of the company unveiling two new, freely available models under an open-source license. OpenAI CEO Sam Altman has confidently described GPT-5 as “the best model in the world.” While this claim might be rooted in pride, independent analysis suggests GPT-5 offers only marginal improvements over leading AI models from competitors like Anthropic, Google DeepMind, and xAI on certain benchmarks, and even slightly underperforms on others.
Despite nuanced benchmark performance, GPT-5 is recognized for its versatility across a broad range of applications, particularly in coding tasks. Crucially, as highlighted by Altman, OpenAI is making significant competitive moves in pricing. “Very happy with the pricing we are able to deliver!” Altman tweeted, signaling a strategic shift aimed at broader adoption and market impact.
The top-tier GPT-5 API is priced at $1.25 per 1 million tokens for input and $10 per 1 million tokens for output, with cached input costing an additional $0.125 per 1 million tokens. This pricing structure closely matches Google’s Gemini 2.5 Pro basic subscription, which is also favored for coding tasks. However, Google implements higher charges when inputs and outputs exceed a significant threshold of 200,000 prompts, potentially making its most intensive users pay more.
OpenAI’s pricing strategy is notably more aggressive than Anthropic’s Claude Opus 4.1, which starts at $15 per 1 million input tokens and $75 per 1 million output tokens. While Anthropic offers substantial discounts for prompt caching and batch processing, OpenAI’s direct per-token cost significantly undercuts the competition, especially for high-volume users.
This competitive pricing has been met with enthusiasm from developers. Simon Willison, an early GPT-5 access user, noted in his review that the pricing is “aggressively competitive with other providers.” Similarly, Matt Shumer, co-founder and CEO of OthersideAI, commented that GPT-5 is “cheaper than GPT-4o, which is fantastic. Intelligence per dollar continues to increase.” On social media platforms like X (formerly Twitter) and forums like Hacker News, users have lauded the pricing, with some calling it “a pricing killer.”
The implications of OpenAI’s aggressive pricing strategy are substantial. If competitors such as Anthropic and Google are compelled to lower their prices in response, it could herald the long-awaited price war in the Large Language Model (LLM) market. Such a development would be a welcome relief for many AI startups, particularly those involved in “vibe-coding” tools, whose economic models are often strained by the high and unpredictable costs of AI model APIs, as previously reported.
The broader AI industry is currently navigating immense infrastructure investments. OpenAI has a $30 billion-per-year contract with Oracle for data center services, while Meta plans to spend up to $72 billion on AI infrastructure in 2025, and Alphabet has allocated $85 billion for capital expenditures in the same year. Amidst these colossal expenditures, the expectation has generally been that costs would rise. However, OpenAI’s strategic pricing move might challenge this trajectory, pressuring the market towards greater affordability.
While it may be too early for startups to fully celebrate the potential reduction in their model API bills, OpenAI’s decisive action this week has undeniably thrown down the gauntlet, putting significant pressure on market pricing. The industry now waits to see if other major players will follow suit.
