OpenAI Releases Full Version of GPT-o1 Inference Model Exclusively for Tier 5+ Developers
Brook.X
OpenAI announced today the release of the full version of the GPT-o1 inference model to select developers. This version boasts superior reasoning capabilities and is better suited for programming development, complex problem-solving, and document drafting scenarios.
Currently, all developers on the API platform have access only to the less powerful GPT-o1 preview versions (such as o1-preview or o1-mini), while the full version of GPT-o1 has been made available to subscribers of ChatGPT Plus and ChatGPT Pro.
The full version of GPT-o1 is exclusively available to certain developers who meet the following criteria:
- Their API platform account is Tier 5 or higher
- The account has been registered for more than one month
- They have spent at least $1,000 on the API platform
Developers who meet these conditions can now access instructions on the OpenAI API platform and, upon confirmation of eligibility, can modify their calls to include this new model.
It's important to note that the GPT-o1 model is not cheap, so for ordinary use cases, it's recommended to continue using the more affordable models to avoid significantly increased expenses.
Additionally, OpenAI has integrated GPT-4o and GPT-4o-mini versions into the OpenAI Realtime API, which is specially built for low-latency voice AI applications, such as creating advanced voice modes similar to ChatGPT.
The Realtime API now also supports WebRTC, potentially enabling more websites to conduct AI voice communications directly with users, ensuring a quality communication experience even on unstable network connections.
