The Tema_Q development team, team zenei, has developed a new importance matrix method called
TaQuants (Tensor-aware Adaptive Quantization).
This model is a TaQuants version of
temaq-org/Tema_Q-X7-Thinking created with TaQuants v3.0.
When combined with the
Tema_Q Agent, it performs agent functions.
While this does not significantly impact the execution of agent tools, there is a risk of model breakdown when generating more than 8k tokens. This is highly likely to be a characteristic specific to the model itself.