Pricing.
ACT
ACT Model has 301 neurons which actually utilizes processing also the neuron based on data compression (MAKES EACH NEURON A INPUT NEURON) which also stores learned reference data in symbolic representation consisting of 6 symbols along with relative weights which makes her turing complete and the entire vast model is stored in a in-memory vector database which only uses index adjustment based on new learning and is static which makes her use less compute but utilizes more memory (RAM).
CHAT LEKH
- 8K Context : $0.06 / 1K tokens
- 32K context : $0.12 / 1K tokens