Flexible pay-as-you-go credit system for LLM token consumption

Flexible Payments

Purchase credits in advance and use them across any service or model. Credits never expire.

Efficient Storage*

Pay only for the storage you use. Indexing improves search speed and accuracy.

*Storage is taken into account in relation to uploaded files and conversations. Users have a free tier of 25GB of storage. Additional charges apply for usage beyond this limit.

Scalable Compute*

Select from a range of compute instances with varying CPU and memory configurations and choose from temporary or permanent instances to suit your computational needs.

*Charges only apply for compute instances that are deployed permanently or if usage exceeds the free tier limits.

Token Consumption and Cost
  • Each time you interact with a language model or agent in Vitral, the system calculates the tokens consumed in real-time and deducts the corresponding cost from your credits. The cost per token depends on the specific model and provider, as shown in the following tables.
  • Users can purchase credits from their Vitral account dashboard. All credits are applicable to any service or model in Vitral and do not expire.
Provider Model Input Cost Output Cost
OpenAI GPT 3.5 $0.00060 $0.01200
OpenAI GPT-4 Turbo $0.00200 $0.00600
OpenAI GPT-4o $0.00100 $0.00300
OpenAI GPT-4o Mini $0.00003 $0.00012
Meta Llama 3.1 405B $0.00060 $0.00060
Meta Llama 3.1 70B $0.00018 $0.00018
Meta Llama 3.1 8B $0.00004 $0.00004
Mixtral Mistral Large 2 $0.00024 $0.00024
Google Gemini 1.5 Pro $0.00070 $0.00210
Google Google Gemini 1.5 Flash $0.00007 $0.00021
Google Gemini 1.0 Pro $0.00010 $0.00030
Anthropic Claude 3.5 Sonnet $0.00060 $0.00300
Anthropic Claude 3 Sonnet $0.00060 $0.00300
Anthropic Claude 3 Opus $0.00300 $0.01500
Anthropic Claude 3 Haiku $0.00005 $0.00025