news

it home reported on september 22 that openai's first models with "reasoning" capabilities, o1 and o1-mini, are now accessible to enterprise and education edition users. users can enter up to 50 messages per week.

according to the introduction,OpenAI the o1 series models have a higher level of comprehension. unlike traditional models, the o1 series models will think deeply before answering questions, generate an internal chain of thinking, and then give answers. therefore, the series models can handle more complex tasks than previous models and are particularly good at solving difficult problems in science, programming, and mathematics.

openai points out that although gpt-4o is still the best choice for most tasks, the o1 series performs better in handling complex, problem-solving-oriented tasks such as research, strategy, programming, mathematics, and science.

however, because the o1 model "will consume more computing resources", openai claims that the series of models will only be available to plus/team/enterprise/edu users in limited quantities in the initial stage.the o1 model can be used up to 50 times per week, while the o1-mini model can be used up to 50 times per day.。

as for the api, the price of access to the relevant model is relatively high, among which the input token fee of o1 is 10 ...15usd (it home note: currently about rmb 107), output word cost is per million 60 usd (currently about rmb 427). when called through the api, the o1 model can be called up to 500 times per minute, and the o1-mini model can be called up to 1000 times per minute.