Medicare’s annual enrollment window is open now and closes December 7. What you can change, and when.
This article is for general educational purposes only and is not insurance or medical advice. We are not affiliated with the U.S. government, Medicare, or any insurer. For official information, visit Medicare.gov or call 1-800-MEDICARE. See our Medical & Financial Disclaimer for details.
CMS has published the file that describes every Medicare Advantage plan you will be able to buy for 2027, and the headline from it is a good one: the average monthly premium is falling again. That is true, and we can confirm it from the file ourselves rather than take it on trust.
But there is a second number in the same file that moved in the opposite direction, and nobody is putting it in a press release. The median 2027 plan caps your in-network spending at $7,100. For 2026 it was $6,500. The monthly price went down about fourteen percent and the price of a bad year went up $600.
We counted both years out of the CMS plan files by the same method, on the same day, so the two columns below are strictly comparable. Here is what actually changed, what it means for the window that opened on October 15, and what the star ratings, published a few days after the plan file, added.
The Number in the Press Release
CMS announced on September 28, 2026, that the average Medicare Advantage premium will fall from $14.37 to $12.00 a month in 2027, a drop of 16.5%. It also expects enrollment to reach 34 million, says more than 99% of beneficiaries will have at least one plan available and 97% will have ten or more, and reports that the Part D premium inside a Medicare Advantage plan falls from $11.32 to $7.00.
None of that is spin. It is the enrollment-weighted average, which means it describes what people are paying rather than what is on the shelf, and on that measure 2027 is cheaper than 2026.
Our own count agrees with the direction, from a different angle. Counting each plan once instead of weighting by how many members it has, the average monthly premium for medical and drug coverage together falls from $19.81 to $17.05, and the share of plans that charge nothing at all per month jumps from 56.4% to 69.7%. That is a real move. Seven plans in ten will now cost you nothing beyond the Part B premium you pay anyway.
So the premium story holds up from both directions. The problem is that the premium is not the number that decides what Medicare Advantage costs you, and it never was.
Both Years, Side by Side, Counted the Same Way
Everything in this table is our own count from the CMS Landscape files, keeping the same plan categories and the same de-duplication for both years. The 2026 column reproduces, to the plan, every figure already published elsewhere on this site.
| Counted from the CMS plan file | 2026 plan year | 2027 plan year | Change |
|---|---|---|---|
| Medicare Advantage plans on sale | 5,553 | 5,522 | −31 |
| Parent companies selling them | 167 | 162 | −5 |
| Counties with at least one plan | 3,147 | 3,131 | −16 |
| Plans with no monthly premium at all | 56.4% | 69.7% | +13.3 points |
| Average monthly premium, medical + drug | $19.81 | $17.05 | −$2.76 |
| Median in-network out-of-pocket limit | $6,500 | $7,100 | +$600 |
| Highest limit any plan carries | $9,250 | $9,850 | +$600 |
| Plans sitting at that highest limit | 25.6% | 28.2% | +2.6 points |
| Plans with an eligibility gate | 1,800 | 1,969 | +169 |
| Plans in the median county | 43 | 42 | −1 |
| Of those, with no eligibility gate | 26 | 23 | −3 |
| Companies in the median county | 7 | 6 | −1 |
Read the table as two halves. Everything about the monthly price got better. Everything about the size of your choice and the size of your worst year got worse.
The $600 That Is Not in Anybody’s Advertising
The out-of-pocket limit is the only figure on a Medicare Advantage plan that tells you how bad a bad year can get. Below it you pay copays; at it you stop paying for covered in-network care. It is the difference between an expensive year and a frightening one, and it is the number the television ads never mention.
Two things happened to it for 2027. The legal ceiling went up: the highest in-network limit any plan carries in the file rises from $9,250 to $9,850, which is the limit set in the CY2027 Rate Announcement that CMS published in April. And more plans went and sat on it — 1,555 of them, up from 1,423.
You might expect that to be concentrated in the special needs plans, which have always clustered at the ceiling for reasons we explain below. It is not. Strip the gated plans out entirely and look only at the plans an ordinary shopper can join:
| Plans with no eligibility gate | 2026 | 2027 |
|---|---|---|
| Plans | 3,753 | 3,553 |
| Median in-network out-of-pocket limit | $5,900 | $6,600 |
| Share sitting at the legal ceiling | 9.0% | 10.8% |
| Share with no Part C premium | 82.0% | 77.2% |
The open market’s median limit rose $700, which is proportionally worse than the market as a whole. So this is not an artifact of the gated plans; it is the ordinary plans getting a little thinner on the one protection that matters most.
Put the two movements together and the trade is explicit. Across the open market the typical plan costs a few dollars less a month — call it $30 or $40 over a year — and exposes you to $700 more in a year when things go wrong. For most people in most years, the premium saving wins. For the minority of people who have the bad year, it does not come close.
That is not an argument against Medicare Advantage. It is an argument for reading the out-of-pocket limit before the premium, which is what our breakdown of what a Medicare Advantage plan actually costs has said all along, and it matters more for 2027 than it did for 2026.
Your Choice Got Slightly Smaller, and Measurably More Gated
The national plan count barely moved — 5,553 to 5,522 — and if that were the whole story it would not be worth a paragraph. The composition moved more than the total.
- Five parent companies left the market, and the median county now has six of them selling rather than seven.
- Gated plans grew while open plans shrank. Special needs plans went from 1,800 to 1,969, and plans anyone can join fell from 3,753 to 3,553. Two hundred ordinary plans left and 169 restricted ones arrived.
- In the median county the ungated choice fell from 26 plans to 23. That is the number that is really yours, and it dropped by more than a tenth.
- Fewer counties have a free ungated plan. In 2026, 2,978 counties out of 3,147 had at least one plan with no gate and no monthly premium. For 2027 it is 2,808 out of 3,131 — about 170 counties fewer.
The growth in gated plans is not bad news in itself, and it is easy to read it wrongly. A special needs plan is restricted, not inferior — and the chronic-condition ones are the cheapest plans in the whole file: for 2027, 92.4% of them charge no monthly premium at all. If you have a long-term illness, Medicaid as well as Medicare, or you need nursing-home-level care, those 1,969 plans are your market too, and our guide to what a special needs plan is and who qualifies goes through the gates one by one. What the shift does mean is that the headline count of plans is drifting further from the count that applies to any individual shopper, which we took apart in detail in how many Medicare Advantage plans there really are.
Where You Live Still Decides More Than the Year Does
A $600 move in the national median is small next to the gap between two states, and that gap survived into 2027 almost unchanged in shape.
| 2027 plan year | Plans on sale | Share with no Part C premium | Median out-of-pocket limit |
|---|---|---|---|
| California | 394 | 82.0% | $3,400 |
| Florida | 560 | 97.7% | $3,975 |
| Nevada | 98 | 96.9% | $4,450 |
| Arizona | 148 | 97.3% | $5,500 |
| Georgia | 161 | 92.5% | $9,850 |
| New Jersey | 71 | 85.9% | $9,850 |
| New York | 224 | 73.2% | $9,850 |
| South Carolina | 97 | 89.7% | $9,850 |
| Massachusetts | 97 | 49.5% | $6,750 |
In Georgia, 92.5% of Medicare Advantage plans charge no medical premium and the median one leaves your spending capped at the highest figure the law allows. In California the median plan caps it at $3,400 — a difference of $6,450 in what a bad year can cost, between two plans that both describe themselves as Medicare Advantage.
Massachusetts sits at the other end of the premium column: barely half its plans are free, and its median limit lands in the middle of the range. And notice what the table does not show, which is any reliable link between its two right-hand columns. Florida, Nevada and Arizona are almost entirely free and among the best protected in the country; Georgia is almost entirely free and the worst. How many plans are free in your state tells you nothing about how well they protect you, which is exactly why you have to read both. Our state-by-state analysis has the full 2027 table for every state, with the 2026 limit alongside for comparison.
One more county-level figure, because it is the one that should decide how much shopping you do: in the median county the gap between the best-protected and the worst-protected plan on sale is $5,450 on the out-of-pocket limit alone. The year-over-year move is $600. The move available to you, this afternoon, by reading the summaries of the plans already on your screen, is nine times bigger.
Three Things a Careful Reader Should Push Back On
CMS says the average premium is $12.00 and you say $17.05. Which is wrong? Neither. CMS weights by enrollment, so its figure describes the premium an average member pays; we count each plan once, so ours describes the premium an average plan charges. Members cluster in the cheaper plans, which is exactly why the weighted figure is lower. Both fell by a similar proportion, and both are stated with their method on our editorial methodology page. Ours also includes the drug premium; the $12.00 headline is the medical premium.
Why a median out-of-pocket limit and not an average? Because the distribution is lumpy: a large block of plans sits exactly on the legal ceiling and drags an average toward it. The median tells you what the middle plan on the shelf looks like, which is the right benchmark when you are judging one plan against its local competition. KFF publishes an enrollment-weighted average for the same quantity and it comes out lower, for the same reason the premium figures differ — we explain that reconciliation in full on the methodology page too.
Where are the star ratings? They were not out when this article first went up: CMS releases the quality ratings separately from the plan file, and the rating column in the 2027 file is empty on all 5,522 plans. CMS published them on October 8, 2026, and we have since recounted every star figure on this site against them. The market as a whole barely moved — 61.2% of rated 2027 plans reach four stars, against 60.5% a year earlier — but individual insurers moved a long way: Humana from 35.5% to 93.6%, Aetna from 72.0% to 48.2%. Our count of which insurers actually score best has the detail, and which plan year each figure describes is set out on the methodology page.
What to Do With This Between October 15 and December 7
The window to act on any of it runs from October 15 to December 7, and December 7 is the date your plan has to have received the request, not the date you mail it. Four things follow from the numbers above, in the order they are worth money.
- Look up your own plan’s 2027 out-of-pocket limit before anything else. If you do nothing, you are renewed into the 2027 edition of your plan, and the median plan’s limit went up $600 this year. The Annual Notice of Change your insurer sent in September has the new figure in it.
- Sort your county’s plans by that limit, not by premium. The spread inside a typical county is $5,450. No amount of premium shopping competes with that.
- Check whether a gate applies to you. More than a third of the plans on sale nationally now have one, and for a chronic condition or dual eligibility the restricted plans are often the cheaper half of your local market. Cheaper is not the same as better protected, though: for 2027 the median chronic-condition plan caps your spending at $6,750, slightly above the $6,600 of an ordinary plan, so read the limit on each one.
- If your plan is one of the ones that left, do not wait. Five companies and two hundred ungated plans came off the market for 2027. A non-renewal gives you a longer window than December 7 and a one-time right to buy a Medigap policy regardless of your health, and that right expires 63 days after your coverage ends.
And one thing not to do: do not treat a falling premium as evidence that your plan got better. For 2027 the two moved in opposite directions, and only one of them is being advertised.
How We Counted This
Every figure attributed to us here is our own count from two CMS Landscape files: the CY2027 file released on September 22, 2026, and revised on September 30, and the CY2026 file from September 2026. Both were processed by identical code on October 2, 2026, which is the only way a year-over-year comparison means anything.
From each file we keep the MA, MA-PD and SNP categories and exclude stand-alone Part D plans, Cost plans and employer group plans, then reduce it to one entry per distinct plan nationally rather than one per county. That gives 5,553 plans in 3,147 counties for 2026 and 5,522 in 3,131 counties for 2027. Gated and ungated come from the file’s own special needs plan indicator. Percentages are shares of plans, not of enrollment, and plans the file marks as having no premium figure are excluded from premium shares rather than counted as zero — which is why the premium shares are worked out over 5,151 plans for 2026 and 5,165 for 2027 rather than over the full totals.
One trap worth naming, because it would have broken the comparison silently: CMS renamed the column that identifies a plan between the two files. In the 2026 file it is ContractPlanSegmentID and in the 2027 file it is Contract Plan Segment ID. Code written for one year does not run on the other, and code that silently skips the de-duplication step would have counted every plan once per county. Our 2026 column reproduces every previously published figure exactly, which is the check that the method survived the rename.
The full procedure, including the reconciliation with KFF’s lower plan count and why both are right, is on our editorial methodology page.
Sources
- Medicare Advantage and Medicare Prescription Drug Programs Expected to Remain Stable in 2027 — Centers for Medicare & Medicaid Services, September 28, 2026. Source for the $14.37 to $12.00 average premium, the 16.5% decline, the 34 million projected enrollment, the access figures of more than 99% and 97%, and the Part D premium figures of $11.32 and $7.00.
- CY2027 Medicare Advantage and Part D Landscape file — Centers for Medicare & Medicaid Services, released September 22, 2026, and revised September 30, 2026. Source for every 2027 figure we counted ourselves: the 5,522 plans, the 162 parent companies, the 3,131 counties, the premium shares and averages, the median and maximum out-of-pocket limits, the 1,969 special needs plans, the per-county figures and the state table.
- CY2026 Medicare Advantage and Part D Landscape file — Centers for Medicare & Medicaid Services, September 2026 version. Source for every 2026 comparison figure in this article, counted by the same method.
- CY2027 Rate Announcement — Centers for Medicare & Medicaid Services, April 6, 2026. The document that sets the maximum out-of-pocket limits Medicare Advantage plans may use in 2027.
- 2027 Star Ratings Data Tables — Centers for Medicare & Medicaid Services, published October 8, 2026. Source for the 2027 star figures quoted in the answer on star ratings, matched to each plan through its contract; the 2026 figures come from the CY2026 file.
Both CMS files were counted on October 2, 2026, and the press release figures checked the same day. Plan availability changes; confirm what is sold at your own address on Medicare.gov before enrolling.
Disclaimer: Plan figures reflect 2026 and 2027 data from the Centers for Medicare & Medicaid Services (CMS) and vary by state, county and plan. Always confirm current details directly with Medicare.gov, 1-800-MEDICARE, or a licensed agent before making a decision.
Unai Sampedro founded and edits Medicare Plan Compass. He is not a licensed insurance agent or broker and takes no commission from any insurer; the site is funded by advertising. Articles are written from CMS and Medicare.gov documents and reviewed each year when new figures are published. Nothing here is personalized advice: the goal is to explain the rules in plain language so readers can ask better questions before they speak to someone who can advise them.