2026.2.9.7 Release Notes

FSD (Supervised) v14.3.1

Full Self-Driving (Supervised) v14.3.1 includes the following changes and improvements:

• Enhanced the Reinforcement Learning (RL) phase used to train the FSD neural network, delivering improvements across a wide variety of driving scenarios.
• Upgraded the neural network vision encoder to better interpret rare and low-visibility situations, strengthen 3D geometry understanding, and broaden traffic sign comprehension.
• Rebuilt the AI compiler and runtime using MLIR, yielding a 20% faster reaction time and accelerating model iteration speed.
• Reduced unnecessary lane biasing and minor tailgating behaviors.
• Made parking spot selection and maneuvering more decisive.
• Improved prediction of the parking location pin, now displayed on the map with a P icon.
• Improved responses to emergency vehicles, school buses, right-of-way violators, and other uncommon vehicles.
• Better handling of small animals by emphasizing harder RL examples and introducing rewards for improved proactive safety.
• Improved traffic light behavior at complex intersections with compound lights, curved roads, and yellow light stopping — informed by hard RL examples sourced from the Tesla fleet.
• Improved management of rare and unusual objects that extend, hang, or lean into the vehicle’s path by leveraging infrequent events from the fleet.
• Improved resilience to temporary system degradations by maintaining control and automatically recovering without driver intervention, reducing unnecessary disengagements.

Upcoming improvements

• Expand reasoning to all behaviors beyond destination handling.
• Add pothole avoidance.
• Improve driver monitoring system sensitivity with better eye gaze tracking, eye wear handling, and higher accuracy in variable lighting conditions.

Requires HW4. Models: S, 3, X, Y, CT.

Full Self-Driving (Supervised)

With driver supervision, Full Self-Driving (Supervised) can operate the vehicle in most situations: it can start from a parked position, perform lane changes, choose forks to follow your navigation route, navigate around other vehicles and objects, make left and right turns, and park at your destination. The driver and any authorized users must remain attentive and exercise additional caution. This feature does not make the vehicle autonomous—do not become complacent.

Full Self-Driving (Supervised) is enabled on your vehicle. To engage it, press the Start Self-Driving button on the UI, or press the right scroll wheel button once. You can disable Full Self-Driving (Supervised) in Autopilot Settings.

Requires HW4. Models: S, 3, X, Y, CT.

Blind Spot Warning While Parked

If you try to open a door while an approaching object is detected in your blind spot (for example, a bicyclist approaching from behind), the blind spot indicator flashes, a chime plays, and the door will not open on the initial button press. Wait briefly, then press the button a second time to override the warning.

Available Worldwide. Models: CT.

Autopilot Name Changes

To align with new naming, the term Autopilot is being replaced with Self-Driving in most areas.

• Under Controls, the Autopilot tab is now labeled Self-Driving.
• Within the Autopilot menu, “Autopilot Features” is now “Self-Driving Features.” This lets the user select between TACC, Autosteer, or Full Self-Driving.

Models: S, 3, X, Y, CT.

Maps Version

The Software menu now shows a simplified map version. Instead of listing region, version, and build, it displays only the region and version, such as NA-2025.44.

Available Worldwide. Models: S, 3, X, Y, CT.

Supercharger Site Maps Expanded

3D Supercharger Site maps were introduced in the 2025 Holiday Update. While that update included 18 Supercharger locations in the United States, the feature is now expanding to Europe. At least two new Supercharger site maps have been added in Belgium, in Brugge and Heusden-Zolder.

These 3D site representations provide an accurate layout coupled with live data.

The layout indicates each parking spot, the charger’s position relative to the space, and highlights handicap- and trailer-friendly stalls, among other details.

Live data shows which stalls are occupied and which Tesla model is charging at each spot. It’ll identify and display a 3D model of the vehicle, and for non-Tesla EVs, it’ll display a generic vehicle model.

Although the 3D models are limited to vehicles with AMD Ryzen infotainment units, anyone can view them by searching for the location in maps.

Available Worldwide. Models: S, 3, X, Y, CT.

Service Mode Improvements

ECU Update Status Panel Updates - Non-CAN ECUs Added

The ECU Update Status panel now includes non-Controller Area Network (CAN) Electronic Control Units (ECUs), such as the Autopilot processors and Telematics Control Unit (TCU).

Models: Model S (2021+), Model 3, Model X (2021+), Model Y, Cybertruck