About Infineon models - Simulation speed and convergence
In general, Infineon models typically run very slowly or returning timestep too small in simulation. This is a well-known issue within the community and is not unique to Qspice. To address these issues, you can use .option fastmath=0 and reduce timestep to a smaller value with .option maxstep=[value] in Qspice.
https://forum.qorvo.com/t/hb-simulation-with-infineons-gan/24205/4
While I couldn’t find a specific reference from Infineon, ROHM, in a similar model, mentioned in its article “New SPICE Models with Improved Simulation Speed for Power Semiconductors Are Released!” that ROHM -L1 is designed for high precision of static characteristics, whereas -L3 offers better convergence and higher speed for evaluating dynamic characteristics. There is no direct comparison to Infineon, as ROHM -L2 seems similar to Infineon -L3, and Infineon does not mention anything about a better convergence model.
https://fscdn.rohm.com/en/products/databook/white_paper/discrete/sic/mosfet/new_spice_model_for_faster_simulation_of_power_semiconductors_wp-e.pdf
I suggest using the Qspice MOSFET Model generator (File > Model Generator > MOSFET) to create a model based on the datasheet. Your simulations will likely run much better and quicker with a native model from Qspice.
This is my unofficial model generator guideline for reference : Model Generators Guide by KSKelvin.pdf