The Omnipod 5 is a tubeless insulin pump that automatically adjusts basal insulin delivery every five minutes based on continuous glucose monitor (CGM) readings. It pairs a small, wearable pod with a CGM sensor and a control algorithm that increases, decreases, or completely pauses insulin to keep blood sugar near a target you set. The result is a closed-loop system that handles much of the minute-to-minute insulin dosing that people with diabetes would otherwise manage manually.
The Core Loop: Pod, CGM, and Algorithm
Three components work together. A disposable pod adheres to your skin (typically on the abdomen, arm, or thigh) and contains both an insulin reservoir and a tiny cannula that sits just under the skin. A compatible CGM sensor, worn separately, measures glucose levels in the interstitial fluid and sends readings to the system. The algorithm lives on the pod itself, so it keeps running even if your phone loses connection temporarily.
Every five minutes, the algorithm receives a fresh glucose reading from the CGM. It compares that value to your programmed target glucose, looks at the direction glucose is trending, and decides how much basal insulin to deliver over the next five-minute window. If glucose is rising, it ramps up delivery. If glucose is dropping, it dials back. When glucose falls below 60 mg/dL, the system pauses insulin delivery entirely to protect against a dangerous low.
This constant micro-adjustment is what separates automated insulin delivery from a traditional pump, which simply delivers a fixed basal rate you program in advance. With Omnipod 5, the basal rate is never truly fixed in automated mode. It shifts continuously based on what your body is actually doing.
Glucose Targets You Can Set
You choose a target glucose level, and the algorithm works to keep you near it. The system allows targets between 110 and 150 mg/dL, adjustable in 10 mg/dL increments. A lower target like 110 mg/dL produces tighter control but slightly more risk of lows, while a higher target gives more cushion. Many users set different targets for different times of day, for example a tighter target during waking hours and a slightly higher one overnight.
Meal Boluses Still Require Input
The automated adjustments handle background insulin needs, but meals are a different story. When you eat, you still need to tell the system how many carbohydrates you’re consuming so it can calculate and deliver a bolus. The built-in bolus calculator factors in your insulin-to-carb ratio, your correction factor, your current CGM reading, and how much insulin is still active from previous doses (often called “insulin on board”). You confirm the suggested dose, and the pod delivers it.
This is the part of the system that still requires the most engagement from you. The algorithm can’t detect food on its own, so forgetting to bolus or underestimating carbs will cause a spike the system then has to chase with increased basal delivery, which is slower and less effective than a properly timed bolus.
The Activity Feature for Exercise and Illness
Exercise, illness, and other situations can make blood sugar less predictable. The Activity feature addresses this by temporarily overriding your programmed target and setting it to 150 mg/dL while also reducing how aggressively the algorithm delivers insulin. You turn it on for a set period of time, and the system becomes more conservative to lower the risk of hypoglycemia.
This is particularly useful before and during aerobic exercise, which tends to drop blood sugar. It can also help during illness when someone isn’t eating normally. The feature is optional and only available in automated mode.
Automated Mode vs. Manual Mode
The system operates in two modes. Automated mode is the full closed-loop experience where the algorithm actively adjusts insulin every five minutes. Manual mode turns off the algorithm and delivers a fixed basal rate that you program, essentially making the Omnipod 5 function like a traditional insulin pump.
The system can drop into a limited automated state if it loses CGM data for an extended period, since the algorithm needs glucose readings to make decisions. Once CGM communication is restored, it returns to full automated mode. This is one reason consistent CGM wear matters: the more data the algorithm has, the better it performs.
Compatible CGM Sensors
Omnipod 5 currently integrates with three CGM sensors: the Dexcom G7, Dexcom G6, and FreeStyle Libre 2 Plus. With the Dexcom sensors, you can manage the system from a compatible iPhone, Android smartphone, or the dedicated Omnipod 5 controller. With the FreeStyle Libre 2 Plus, management runs through the Omnipod 5 app on the provided controller only. Dexcom is in the process of discontinuing the G6, so new users will typically start with the G7.
The Pod Itself
Each pod is disposable and worn for up to 72 hours (three days) before replacement. It’s waterproof with an IP28 rating, meaning it can be submerged to about 7.6 meters (roughly 25 feet) for up to 60 minutes. You don’t need to remove it for swimming or showering. The pod is small, roughly the size of a few stacked credit cards, and attaches with an adhesive backing. There are no tubes connecting it to anything else on your body.
Learning Period and Adaptation
When you first start using Omnipod 5 in automated mode, the algorithm begins with your existing insulin settings (basal rates, correction factors, carb ratios) and adapts over time. It uses your total daily insulin history to refine its predictions, so the first several days may not represent its best performance. Most users and clinicians report that the system gets noticeably better after the first week or two as it accumulates data specific to your body.
How Well It Works in Practice
In a retrospective study of youth with type 1 diabetes, those who started on Omnipod 5 saw their time in range (the percentage of the day spent between 70 and 180 mg/dL) increase by about 11.8 percentage points over 90 days. Users went from an average of 51.6% time in range at baseline to 64.1%. That improvement is clinically meaningful: each additional percentage point of time in range corresponds to measurable reductions in long-term complication risk. The recommended target for most people with type 1 diabetes is at least 70% time in range, so many users still benefit from optimizing meal boluses and settings alongside the automation.
The system’s biggest impact tends to be overnight, when it can make small adjustments while you sleep. Many users report waking up with more stable glucose levels than they achieved with manual pumping or injections, since the algorithm keeps working through hours when manual management isn’t possible.