OpenAI launches GPT-4.5 Orion: the most expensive AI model

Post Thumbnail

OpenAI introduced GPT-4.5 under the code name Orion — the largest artificial intelligence model in the company’s history. The new product, which required record computing power and training data volume, demonstrates both impressive achievements and unexpected limitations of the traditional approach to AI development.

Access to GPT-4.5 has already been granted to ChatGPT Pro subscribers ($200 per month) and developers using paid API plans. ChatGPT Plus and ChatGPT Team users will be able to try the model next week. However, the cost of using GPT-4.5 through the API turned out to be unprecedentedly high: $75 per million incoming tokens (about 750,000 words) and $150 per million outgoing tokens, which is 30-15 times more expensive than GPT-4.

According to testing, the new model demonstrates a deeper understanding of the world and increased emotional intelligence. GPT-4.5 beats competitors in 57-63% of cases, especially in professional queries. Among the key improvements are increased creativity, the ability to provide more relevant recommendations in subjective areas, and improved explanation of complex scientific concepts.

The model has particularly excelled in specific tasks such as generating SVG code from sketches, developing game strategies, and checking texts for contradictions. Users note more “human-like” communication and improved “sense of taste” of GPT-4.5.

However, OpenAI acknowledges that the model has not met all expectations. On some tests, it falls short of new “reasoning” models from DeepSeek, Anthropic, and OpenAI’s own developments. This may indicate that the traditional approach to AI training, based on increasing computational power and data volume, is beginning to reach its limit of effectiveness.

Due to high operational costs, OpenAI is considering discontinuing support for GPT-4.5 in its API in the long term. This decision could significantly impact the future development of artificial intelligence technologies and force the industry to seek new approaches to model training.

Почитать из последнего
UBTech will send Walker S2 robots to serve on China's border for $37 million
Chinese company UBTech won a contract for $37 million. And will send humanoid robots Walker S2 to serve on China's border with Vietnam. South China Morning Post reports that the robots will interact with tourists and staff, perform logistics operations, inspect cargo and patrol the area. And characteristically — they can independently change their battery.
Anthropic accidentally revealed an internal document about Claude's "soul"
Anthropic accidentally revealed the "soul" of artificial intelligence to a user. And this is not a metaphor. This is a quite specific internal document.
Jensen Huang ordered Nvidia employees to use AI everywhere
Jensen Huang announced total mobilization under the banner of artificial intelligence inside Nvidia. And this is no longer a recommendation. This is a requirement.
AI chatbots generate content that exacerbates eating disorders
A joint study by Stanford University and the Center for Democracy and Technology showed a disturbing picture. Chatbots with artificial intelligence pose a serious risk to people with eating disorders. Scientists warn that neural networks hand out harmful advice about diets. They suggest ways to hide the disorder and generate "inspiring weight loss content" that worsens the problem.
OpenAGI released the Lux model that overtakes Google and OpenAI
Startup OpenAGI released the Lux model for computer control and claims this is a breakthrough. According to benchmarks, the model overtakes analogues from Google, OpenAI and Anthropic by a whole generation. Moreover, it works faster. About 1 second per step instead of 3 seconds for competitors. And 10 times cheaper in cost per processing 1 token.