OpenAI announces GPT-4.5
OpenAI announced on February 27, 2025, the launch of GPT-4.5, described by the company as its largest and most capable language model to date. The new model is designed to offer better performance in reasoning, accuracy, and reduction of hallucinations, according to an official statement.
Initial availability
GPT-4.5 was initially released to ChatGPT Pro subscribers, with plans to expand to other subscription tiers in the coming days. The company did not disclose details about pricing or an exact timeline for release to other users.
Capabilities and improvements
According to OpenAI, GPT-4.5 represents a significant advancement over previous models, especially in complex reasoning ability and reduction of fabricated responses (hallucinations). The company claims the model is more accurate and reliable, although it did not provide specific performance metrics.
Launch context
The announcement comes amid increasing competition in the artificial intelligence sector, with companies like Google, Anthropic, and Meta launching their own advanced models. OpenAI seeks to maintain its leadership with GPT-4.5, which is expected to compete directly with Gemini 2.0, Claude 3, and LLaMA 3.
Next steps
OpenAI plans to release GPT-4.5 to all ChatGPT subscribers in the coming weeks, starting with the highest tiers. The company also indicated it will continue improving the model based on user feedback. There is no forecast for availability via API or for free users.