Q: Hey Felipe, you are definatley on to something amazing here! Triplo can be used for SO MANY use cases!
Something that I am confused about is that do the included tokens mean 1) we get to use AI without adding our own API key up to these tokens and then we can add our own API key for further use; OR 2) is it that we have to add our API key to start with? If this is the option 1, then are all the models available within the app? And the number of words we can generate and actions we can perform will depend on the model we use and the length of input prompt and context?
Thanks
Jay
Felipe_TriploAI
Nov 5, 2024A: Hello Jay! It’s truly wonderful to connect with you. Let me address your questions.
1. We provide our users with a generous monthly allowance of tokens per device. Specifically, you can process up to 2 million tokens each month, which is roughly equivalent to about 1.5 million words or 17 complete romance novels per device. You can choose from over 15 models within this allowance, including:
- GPT-4o-mini 128k
- Dolphin Mixtral 8x7B 32k
- Google Gemini Pro 1.0 131k
- Meta Llama-2 13B 4k
- Nous Llama-2 13B 4k
- Meta Llama-3 8B Instruct 8k
- Meta Llama-3.1 8B Instruct 131k
- Meta Llama-3.1 70B Instruct 100k
- Mistral-7B 8k
- Mistral Tiny 32k
- MythoMax 13B 4k
- MythoMax 13B 8k
- OpenChat 3.5 7B 8k
- Toppy M 7B 32k
- Mistral Open Orca 7B 8k
- MythoMist 7B 32k
- Open Hermes 2.5 7B 4k
- PPLX Online 7B 4k
- Yi 6B 4k
Based on my experience, these models will meet the needs of 99% of regular users.
If you wish to access premium models (like O1, Claude, etc.), you can bring your own API key, commonly referred to as BYOK (Bring Your Own Key). This will grant you access to additional models, including:
- GPT-3.5 16k
- GPT-4o 128k
- GPT-4 128k
- GPT-4 8k
- GPT-4 32k
- Claude-1 100k
- Claude-2 100k
- Claude-3 variants (Haiku, Opus, Sonnet) with up to 200k tokens
- Google PaLM-2 8k
- Google Gemini Pro 1.5 2.8M
- And many more...
I hope this clarifies everything for you.
Take care,
Felipe