So, look at the above picture. Do you see the Gemini Experimental 1206 (a.k.a
Gemini-exp-1206
) with the new mark on it? Well, in the Gemini API Changelog, they don’t update that, but we can use it in Python ( I don’t know if node.js works or not) by using the Model ID to use it.
So, this is the short video for the testing from the Gemini Experimental 1206 using Python (I use .env to censor my API key):
Drive Video
Hi @qvu217720,
Thanks for flagging, I will convey this with the team.
Hi,
In the Gemini API one could use the model list regularly to see which new models have been made available to your account. Using the the actual model identifier, eg. models/gemini-exp-1206
, is normal regarding the Gemini API and also described like this.
Please note, it already happened that models exposed in Google AI Studio were not available in the Gemini API and vice versa at the same time.
It might be that the documentation is lagging behind a little bit though - as usual.
Cheers, JoKi
Ok, but in my code that uses Gemini-exp-1206, it ran successfully.
Yes, same here.
There can lag between those three players - Google AI Studio, Gemini API and documentation - in any direction, and there had been in the past.
As for the API always check the list of available models and you’re on the safe side. I have a REST collection of requests, like this
#@apiKey={{$dotenv GOOGLE_API_KEY}}
# Get a list of available models.
GET https://generativelanguage.googleapis.com/v1beta/models
x-goog-api-key: {{apiKey}}
Which I run regularly in order to check if there are new models available in my account.