Tesla’s Full Self-Driving (Supervised) v14.3.2 arrives via software update 2026.2.9.8, delivering neural network upgrades, 20% faster reactions, and unified AI models across key features. This rollout targets HW4-equipped vehicles like Model S, 3, X, Y, and Cybertruck, with early drives showing real-world gains in parking and edge cases.
Core neural network upgrades
Reinforcement Learning training improved for diverse scenarios, while the vision encoder now handles rare events, low visibility, 3D geometry, and more traffic signs. A ground-up AI compiler rewrite using MLIR boosts reaction speed by 20% and iteration efficiency, cutting lane bias, tailgating, and disengagements via auto-recovery.
Model unification across features
A key release note highlight, Actually Smart Summon, FSD, and Robotaxi now share a unified neural network model for more capable, reliable behavior. This ensures consistent performance in summoning, driving, and robotaxi ops, reducing edge-case inconsistencies like hesitant parking.
Parking and maneuvering gains
Spot selection grew more decisive with better maneuvering and map pins showing predicted “P” locations. Smart Summon benefits directly from the unified model, appearing smoother in user tests.
Edge case and safety enhancements
Better responses to emergency vehicles, school buses, violators, small animals, complex lights/intersections, and path-obstructing objects stem from fleet-sourced RL training. Temporary faults now recover automatically; post-intervention reason selection helps Tesla refine the system.
Upcoming enhancements
Plans include full-behavior reasoning, pothole avoidance, and advanced driver monitoring with eye gaze tracking for glasses and varied lighting.
Additional 2026.2.9.8 features
Autopilot rebrands to Self-Driving, with parked blind spot alerts, expanded Supercharger maps, and service tweaks. Rollout started around April 22, 2026, building FSD v14 reliability.
Official release notes: FSD (Supervised) v14.3.2
Included in 2026.2.9.8
Full Self-Driving (Supervised) v14.3.2 includes:
- Upgraded the Reinforcement Learning (RL) stage of training the FSD neural network, resulting in improvements in a wide variety of driving scenarios.
- Upgraded the neural network vision encoder, improving understanding in rare and low-visibility scenarios, strengthening 3D geometry understanding, and expanding traffic sign understanding.
- Rewrote the AI compiler and runtime from the ground up with MLIR, resulting in 20% faster reaction time and improving model iteration speed.
- Mitigated unnecessary lane biasing and minor tailgating behaviors.
- Increased decisiveness of parking spot selection and maneuvering.
- Improved parking location pin prediction, now shown on a map with a P icon.
- Enhanced response to emergency vehicles, school buses, right-of-way violators, and other rare vehicles.
- Improved handling of small animals by focusing RL training on harder examples and adding rewards for better proactive safety.
- Improved traffic light handling at complex intersections with compound lights, curved roads, and yellow light stopping — driven by training on hard RL examples sourced from the Tesla fleet.
- Improved handling for rare and unusual objects extending, hanging, or leaning into the vehicle path by sourcing infrequent events from the fleet.
- Improved handling of temporary system degradations by maintaining control and automatically recovering without driver intervention, reducing unnecessary disengagements.
- Help Tesla improve Self-Driving by selecting an intervention reason on the main screen after taking over.
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.
We’ve launched an official Tesla Owners Online Forum. Join early and claim your "Founding Fuel" badge.
