Tesla Software Update 2025.45.10 Rolls Out FSD v14.2.2.5 — Here's Everything That Changed

Published February 21, 2026  ·  5 min read

Tesla has kicked off a new rollout wave, pushing software version 2025.45.10 — and with it, FSD (Supervised) v14.2.2.5 — to a growing pool of vehicles. Teslascope first flagged the wave hitting cars within a 30-minute window on February 18, with earlier sightings of the build traced back to February 15–16, 2026. The deployment is currently a slow, controlled release targeting Hardware 4 (HW4)-equipped vehicles, including the Model Y, Model 3, and Cybertruck, with broader waves expected to follow.

The v14.2.2 series has been one of the more substantive FSD revision cycles in recent memory, and 14.2.2.5 carries forward and refines that work. Here is a full breakdown of what this update actually delivers.

📊 Key Figures

What's Actually New in FSD v14.2.2.5

According to verified release notes from multiple tracking sources, the v14.2.2 series — now refined in this .5 point release — includes the following substantive changes:

🧠 Neural Network Vision Upgrades

Tesla has upgraded the core neural network vision encoder underpinning FSD. The improvements focus on more accurate detection of emergency vehicles (police cars, fire trucks, ambulances), better identification of road obstacles, and the ability to read and respond to human gestures — such as a pedestrian waving a vehicle through. These are not minor tweaks; they represent meaningful changes to how the car perceives and reasons about its environment in complex real-world scenarios.

🚨 Emergency Vehicle Handling

Building on the vision improvements, FSD v14.2.2.5 adds explicit logic for pulling over and yielding to emergency vehicles. This is a compliance and safety feature that has been a long-standing gap in earlier FSD versions. The system should now reliably recognize active emergency responses and respond correctly without driver intervention being required to manage the situation.

🅿️ Arrival Options: Choose How You're Dropped Off

A genuinely useful quality-of-life addition: Arrival Options now lets you specify your preferred parking endpoint before you arrive. Options include Parking Lot, Street, Driveway, Parking Garage, and Curbside. For FSD users in urban environments or those dropping off passengers, this removes the ambiguity of where FSD decides to stop at the end of a route.

🗺️ Navigation Integrated Into the Vision Network

Perhaps the most architecturally significant change: Tesla has integrated navigation and routing data directly into the vision-based neural network. Previously, navigation instructions were layered on top of the driving stack as separate inputs. Now, the car can respond in real time to blocked roads and unexpected detours at the neural network level — enabling more adaptive, context-aware route management rather than rule-based overrides.

🐢🔥 New Speed Profiles: SLOTH and MAD MAX

Tesla has added two new named Speed Profiles that sit at opposite ends of the driving style spectrum. SLOTH opts for lower speeds and conservative lane choices — useful in high-traffic areas, bad weather, or for new FSD users who prefer a gentler experience. MAD MAX prioritizes higher speeds and more frequent lane changes for experienced users who want FSD to keep pace with faster traffic flows. These sit alongside the existing profiles and are user-selectable.

🚧 Obstacle and Gate Handling

The update improves handling of both static and dynamic gates — such as those found at parking garages or gated communities — and adds better offsetting logic for road debris. The car should now navigate around stationary objects in its path more smoothly rather than hesitating or requiring intervention.

🔄 Scenario Refinements

Tesla lists refinements across several common FSD challenge scenarios, including unprotected turns, lane changes, vehicle cut-ins, and school bus interactions. These are the types of edge cases that have historically generated the most disengagements, and iterative improvements here have a disproportionate impact on real-world usability.

📷 Windshield Residue Alert

A practical new alert warns drivers when residue build-up on the interior windshield may be affecting front camera visibility. This is a proactive safety measure — camera occlusion from interior film (a common issue in cold or humid climates) can degrade FSD perception, and the system will now flag it directly.

🖥️ UI and Fault Recovery Updates

The interface gains a 'Self-Driving stats' section under Controls > Autopilot, and owners can now initiate FSD with a touchscreen tap in addition to the stalk. On the reliability side, FSD's ability to manage system faults and recover from degraded operation states has been improved — reducing the likelihood of abrupt disengagements in marginal conditions.

🔭 The BASENOR Take

📰 Deep Dive

The v14.2.2.x series has been incrementally building toward something Tesla has been pursuing for years: a FSD stack where navigation context, vision, and decision-making are not separate modules bolted together, but a unified system. The integration of routing into the vision neural network in this release is a concrete step in that direction. When a road is blocked, the car does not wait for the navigation layer to catch up — it adapts in real time at the perception level. That is a meaningful architectural shift, not a feature checkbox.

The new speed profiles — SLOTH and MAD MAX — are worth paying attention to beyond their memorable names. They reflect Tesla acknowledging that FSD's default driving style is not universally appropriate. A single default tuning cannot serve both a nervous first-time FSD user on suburban streets and an experienced owner commuting on a fast-moving freeway. Giving owners explicit control over the risk/comfort trade-off is the right call, and the naming makes the options genuinely intuitive rather than buried in a settings menu.

The windshield residue alert is quietly one of the most practically useful additions. Interior windshield film is an underappreciated source of camera degradation — especially in winter climates where temperature differentials cause persistent fogging and haze to accumulate on the inside of the glass. Owners who have been experiencing inconsistent FSD behavior without an obvious cause may find that this alert surfaces a culprit they were not aware of.

One important caveat for most owners: this is still a limited, staged rollout targeting HW4 vehicles first. If your car has not received 2025.45.10 yet, that is expected — not a sign that something is wrong with your vehicle or account. Additional waves are expected to follow as Tesla validates the build at scale. If you are eager to push the update, keeping your vehicle connected to Wi-Fi and checking the Software section of your Tesla app regularly is the best way to catch it as soon as your vehicle is selected.