OpenAI has officially launched GPT-4.5, the highly anticipated AI model known by its codename, Orion. As the company’s largest model to date, GPT-4.5 has been developed with significantly greater computational resources and training data than its predecessors.
Despite its scale, OpenAI clarifies in a recently released white paper that GPT-4.5 is not considered a frontier model.
Starting Thursday, subscribers to OpenAI’s premium ChatGPT Pro plan, which costs $200 per month, will gain early access to GPT-4.5 within ChatGPT as part of a research preview. Developers using OpenAI’s paid API tiers will also be able to access the model immediately. Other ChatGPT users, including those subscribed to ChatGPT Plus and ChatGPT Team, can expect to receive access sometime next week, according to an OpenAI spokesperson speaking to TechCrunch.
The release of Orion has been eagerly awaited by industry experts, as it serves as a critical test for the continued effectiveness of traditional AI training methodologies. Like its predecessors—GPT-4, GPT-3, GPT-2, and GPT-1—GPT-4.5 was developed using an approach that dramatically scales up computational power and data during the unsupervised learning phase of training.
Historically, each new generation of GPT has seen dramatic performance improvements across multiple domains, including mathematics, writing, and coding. OpenAI claims that GPT-4.5 boasts “deeper world knowledge” and “higher emotional intelligence.” However, signs suggest that the performance gains from scaling are beginning to plateau. In key AI benchmarks, GPT-4.5 trails some of the latest reasoning models from competitors such as DeepSeek, Anthropic, and even OpenAI’s own research efforts.
Running GPT-4.5 is also extremely costly. OpenAI acknowledges this challenge, stating that it is evaluating whether the model’s long-term availability in the API is viable. The pricing for GPT-4.5’s API access is set at $75 per million input tokens (about 750,000 words) and $150 per million output tokens—significantly higher than GPT-4o, which costs just $2.50 per million input tokens and $10 per million output tokens.
“We’re releasing GPT-4.5 as a research preview to better understand its strengths and limitations,” OpenAI stated in a blog post shared with TechCrunch. “We are still discovering its capabilities and are eager to see how users apply it in unexpected ways.”
Mixed Performance Results
OpenAI has emphasized that GPT-4.5 is not intended to replace GPT-4o, the company’s primary model for powering ChatGPT and the API. While GPT-4.5 does support features such as file and image uploads and ChatGPT’s canvas tool, it currently lacks real-time voice interaction, which is available in GPT-4o.
On the performance front, GPT-4.5 outshines many models, including GPT-4o, in certain areas. On OpenAI’s SimpleQA benchmark, which measures factual accuracy, GPT-4.5 surpasses GPT-4o and OpenAI’s reasoning models, o1 and o3-mini. OpenAI also reports that GPT-4.5 hallucinates less often than most AI models, making it theoretically more reliable.
We have helped 20+ companies in industries like Finance, Transportation, Health, Tourism, Events, Education, Sports.
