OpenAI plans to phase out GPT-4.5, its largest-ever AI model, from API availability. The AI giant announced on Monday that GPT-4.5 will no longer be accessible through its API starting July 14, urging developers to adopt GPT-4.1 as the new standard.
OpenAI plans to phase out GPT-4.5 just months after its February release, marking a surprisingly short life span for a model that was once positioned as a major leap forward in AI capabilities. The company confirmed the decision via email, noting that GPT-4.1 not only matches but often surpasses GPT-4.5’s performance — at a much lower operational cost.
Why OpenAI Is Retiring GPT-4.5 So Soon
According to OpenAI, the reason behind the shift is clear: performance and cost-efficiency. While GPT-4.5 was an impressive technical feat — trained on more compute power and data than any model before it — the financial burden of running it proved unsustainable for long-term use in OpenAI’s API.
“[GPT-4.1] offers similar or improved performance than GPT-4.5 in key areas at a much lower cost,” an OpenAI spokesperson explained. “We will be deprecating GPT-4.5 to prioritize building future models.”
It’s important to note that while OpenAI plans to phase out GPT-4.5 from the API, the model will still be available in ChatGPT’s research preview for paying subscribers.
GPT-4.5: A Powerful but Costly Experiment
Codenamed “Orion,” GPT-4.5 was built to push the limits of natural language processing and code generation. The model brought noticeable enhancements over GPT-4o, especially in writing, logical reasoning, and argument structuring.
Despite its size and training depth, however, GPT-4.5 struggled to reach “frontier-level” status on several industry-standard AI benchmarks, reducing its long-term appeal for developers seeking cutting-edge performance.
Adding to the issue was its hefty price tag: GPT-4.5 cost $75 per million input tokens and $150 per million output tokens — making it one of the most expensive models OpenAI has ever offered for public API use.
GPT-4.1 Takes Center Stage
With OpenAI planning to phase out GPT-4.5, the company is positioning GPT-4.1 as the more efficient and capable successor. Launched the same day as the announcement, GPT-4.1 is optimized for both coding and real-world software engineering tasks and offers a significantly lower cost structure for developers.
This pivot also signals OpenAI’s increasing focus on making its AI models both more scalable and accessible, particularly as competition heats up from Google’s Gemini and Anthropic’s Claude AI.
As OpenAI plans to phase out GPT-4.5, the message to developers is clear: get ready to transition before July 14 — and prepare for even more advanced models on the horizon.
Get the Latest AI News on AI Content Minds Blog