OpenAI New Model, GPT-4.5, Accessible to ChatGPT Plus Users
OpenAI has recently expanded access to its latest AI model, GPT-4.5, specifically for ChatGPT Plus subscribers. Initially available to users on the $200-a-month ChatGPT Pro plan, the model is now gradually reaching a wider audience.
According to OpenAI, the rollout of GPT-4.5 is expected to take 1 to 3 days, with potential adjustments to rate limits based on user demand over time.
Advanced Capabilities of GPT-4.5
GPT-4.5 represents OpenAI's most advanced model yet, trained with significantly more data and computational power compared to its predecessors. This has resulted in a broader knowledge base and enhanced emotional intelligence. While it marks progress in various aspects, it may not necessarily surpass all other existing AI models.
Competing AI models from companies like DeepSeek and Anthropic have demonstrated stronger reasoning capabilities in specific tests, prompting discussion on whether OpenAI's emphasis on scaling up truly translates to superior performance.
Cost and Performance Considerations
Despite its impressive capabilities, GPT-4.5 comes with a substantial cost. OpenAI acknowledges the high operational expenses associated with running the model, leading them to contemplate its long-term availability in their API.
Currently, OpenAI charges $75 per million tokens for input and $150 per million tokens for output, making it significantly more expensive than GPT-4. Input costs are 30 times higher, while output costs are 15 times higher than those of GPT-4.
Enhancements and Advantages
Despite the costs, GPT-4.5 boasts notable improvements. It exhibits reduced instances of "hallucinations," minimizing the generation of incorrect or misleading information. The model excels in rhetorical tasks, showcasing advanced reasoning and negotiation skills.
OpenAI's internal evaluations highlight GPT-4.5's ability to persuade other AI models to perform tasks effectively, underscoring its enhanced capabilities.
Future Outlook
The future of GPT-4.5 remains uncertain as OpenAI deliberates on the model's continued presence in their API given the high operational costs. Users are encouraged to explore its enhanced features, but long-term availability is not guaranteed.
OpenAI is carefully evaluating performance metrics, user demand trends, and operational costs to determine the optimal path forward regarding GPT-4.5.
