FSD Version History › FSD 13.0

Tesla FSD 13.0

Released

Overview

FSD V13 is a major architecture update featuring a significantly larger neural network model and expanded training data. Elon Musk described it as '5-6× better' than V12.5.

What's new in FSD 13.0

  • New, significantly larger end-to-end neural network model
  • Expanded training dataset — billions of miles of fleet data
  • Improved handling of rare edge cases and novel traffic scenarios
  • Better performance in adverse lighting conditions (rain, glare, night)
  • More human-like decision-making at complex intersections
  • Improved detection of small objects and debris in the road

Known issues

  • Some initial reports of over-cautious behavior on familiar routes
  • Higher compute load may affect older HW3 performance

Community feedback

V13 is widely regarded as a step-change in FSD capability. Independent testers report noticeably fewer interventions per 100 miles compared to V12.5, especially in complex urban scenarios.

How does Tesla FSD 13.0 compare?

See how Tesla FSD stacks up against Waymo, Mobileye, and other autonomous driving systems.

View FSD vs Waymo comparison →

Frequently asked questions

What is new in FSD 13.0?

New, significantly larger end-to-end neural network model. Expanded training dataset — billions of miles of fleet data. Improved handling of rare edge cases and novel traffic scenarios. Better performance in adverse lighting conditions (rain, glare, night). More human-like decision-making at complex intersections. Improved detection of small objects and debris in the road

What are the known issues in FSD 13.0?

Some initial reports of over-cautious behavior on familiar routes. Higher compute load may affect older HW3 performance

What do FSD 13.0 users say?

V13 is widely regarded as a step-change in FSD capability. Independent testers report noticeably fewer interventions per 100 miles compared to V12.5, especially in complex urban scenarios.

Related autonomous driving research

Explore the academic papers behind FSD's end-to-end neural network approach: