Oh, sweet: both o3-mini low and high support integrated web search. No integrated web search with o1.
I prefer, for philosophical reasons, open weight and open process/science models, but OpenAI has done a very good job at productizing ChatGPT. I also use their 4o-mini API because it is cheap and compares well to using open models on Groq Cloud. I really love running local models with Ollama but the API venders keep the price so low that I understand most people not wanting the hasssle if running Deepseek-R, etc., locally.
I prefer, for philosophical reasons, open weight and open process/science models, but OpenAI has done a very good job at productizing ChatGPT. I also use their 4o-mini API because it is cheap and compares well to using open models on Groq Cloud. I really love running local models with Ollama but the API venders keep the price so low that I understand most people not wanting the hasssle if running Deepseek-R, etc., locally.