#Pricing for gemini at Generative Language API
1 messages · Page 1 of 1 (latest)
Hi @wanton gulch. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
I'm sorry, but as an AI developed by OpenAI, I don't have information about Google's Generative Language API or its pricing. My knowledge is limited to the information provided in the extracts, which are all related to OpenAI's products and services. I recommend checking Google's official documentation or reaching out to their support for accurate information.
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
@wanton gulch kapa.ai is still learning and improving, please let us know how it did by reacting below
can i ask the pressing questions like this too
Hello @wanton gulch! I have tested it since beggining. Its free. You can use it on Python and Jupyter Notebook( or any other enviroment). No Need to pay. Just create an API key to use as local host. Hope that this help!
Gemini Pro is generally free across all products. It will be Gemini Ultra that will be a premium version of Bard, and I assume part of their cloud products
True...i am using it to create some models. But Even the Ultra( is available) have too many bugs...But, for some kind of subjects, have good answers. i.e: to get data from Scoppus, Scince Direct. Well...with new GPT resources, make no sense use Gemini for simple tasks.
You have early access? That's cool. How does it compare to GPT-4? Their announcement page very gladly showed off where Bard Ultra beat GPT-4 in tests.
I had early access. Since the model I use is private, it's quite different (and not intuitive) compared to the public model, which is generalist. Regarding the difference with GPT-4, it becomes more visible in terms of what I would like and the solution for real-time scientific answers (it denies it, but it accesses it, hahahaha). Modifying it using other models allows the creation of DataFrames, predictive modeling that we can't change through GPT.
From the tests I've done and the model they provide, it focuses not on assistance but on scientific resolution. However, I always use GPT and Gemini to confront the information. Then try a simple command: "mention the 20 most relevant articles on a complex topic, send the metadata "(it does this fiercely) and in real-time (even if it denies it). This bug of not knowing in which year it was done, if it's the Bard, etc., ends up being interesting. If you want, I've already encapsulated it in Python (or if you want the notebook, I can send it as well). Just be careful with the API key and the bizarre information it sends.
No, I don't want to use someone else's API key, but thanks!
No..i mean you create your API. On GitHub have all these steps.
It is free as listed on the web.
However, what do you mean by hosting it myself? Is the model opensource? Any details on how to do this?
Hello ! Well, yes. All free and you are the owner of what you did( not google). You can host it on your computer, IoT device, Android etc. Languages( Python, Node,Go etc). The models are opensource and can be customized( with examples to create your own model). On the image above, you just need to login with your Google account and create an API key. You can make many API keys. " The G Colab or Vertex " is useful to people that want see as GPT and in the cloud to avoid to create the codes( as GPT is). But in the Cloud is the public Model( more stable, generalist, can´t be customized, less hallucination)
The private one, is much more powerful, highly specialized for what you need, and it doesn't share your training data. (It generates a lot of hallucination for simple questions, less stable, bugs and need to code)
On GitHub, have a lot of things. etc. I can´t put the sites here( blocking).
Hope that it help! Good day!
I have used the model with api key. Its not really "hosted" on my own machine though.
Is there like a model I can download and then prompt it locally without having to send a api request to google?