• Announcement of GPT-4.1 Family: On Monday, OpenAI announced the GPT-4.1 model family with a 1 million token context window. It includes GPT‑4.1, GPT‑4.1 mini, and GPT‑4.1 nano. These models outperform GPT-4o in some areas and will only be available through the developer API, not in ChatGPT.
• Context Window and Comparison: The 1 million token context window allows these models to handle a large amount of text. It puts OpenAI's context windows on par with Google's Gemini models. GPT-4.1 is compared to GPT-4.5 Preview, which will be retired in the API. GPT-4.1 scores better on coding benchmarks but GPT-4.5 performs better on academic knowledge tests and vision-related tasks.
• Pricing and Cost: The new models have lower prices than their predecessors. GPT-4.1 costs $2 per million tokens for input and $8 per million tokens for output, with GPT-4.1 mini and nano having even lower prices. GPT-4.5 was very expensive through the API.
• API-Only Strategy: OpenAI offers these advanced models only through the API to provide developers with specific and consistent models. ChatGPT users receive a constantly evolving GPT-4o model that absorbs capabilities over time. This creates a two-track system but makes naming more confusing.