The 2025 CMMI Scorecard
Ten Models in One Year
Between March and December 2025, the CMS Innovation Center cancelled four models, halted two before they launched, announced nine new models, and redesigned one active model. That pace of activity, concentrated in ten months, is without precedent in CMMI’s fifteen-year history. The result is a fundamentally different Innovation Center than the one that entered 2025, operating under a different strategic framework, targeting different policy objectives, and testing a different theory of how Medicare payment reform generates savings.
This article inventories the full 2025 portfolio, assesses its internal logic, identifies what the portfolio does not include, and evaluates what the aggregate design reveals about where CMMI is heading.
The Cancellations and the Signal They Sent#
On March 12, 2025, CMS announced the early termination of four active models: Primary Care First, Making Care Primary, the ESRD Treatment Choices Model, and the Maryland Total Cost of Care Model (replaced by AHEAD). Two additional models, the Medicare $2 Drug List and the Accelerating Clinical Evidence model, were halted before implementation. CMS cited the CBO’s 2023 finding that CMMI’s cumulative activities had increased net federal spending by $5.4 billion and framed the cancellations as a reset toward models that generate certifiable savings rather than expanding coverage or testing delivery innovation without demonstrated fiscal return.
The cancellations were not random. They targeted voluntary primary care models (PCF, MCP) whose evaluations showed limited or no net savings, a disease-specific model (ESRD-ETC) with mixed results, and a state model (Maryland TCOC) being replaced by AHEAD’s expanded framework. The pattern established the administration’s threshold: models that do not produce savings CMS’s actuary can certify are not worth continuing. That threshold shaped everything that followed.
The 2025 Portfolio#
The nine new models and one major redesign announced between May and December 2025 can be organized along two dimensions: the type of accountability they impose and whether participation is mandatory or voluntary.
Mandatory models with direct financial consequences:
WISeR (Wasteful and Inappropriate Service Reduction), announced May 2025, introduces AI-powered prior authorization to Original Medicare FFS for the first time. It operates in six states across four MAC jurisdictions, targeting services with high rates of inappropriate utilization including skin substitutes, spinal procedures, nerve stimulators, and knee arthroscopy. Contractors face financial incentives tied to denial accuracy and savings generation. WISeR is the administration’s answer to the $5.8 billion in annual Medicare FFS spending CMS identifies as unnecessary or inappropriate.
ASM (Ambulatory Specialty Model), proposed July 2025 and finalized October 2025, imposes mandatory two-sided risk on specialists treating heart failure and low back pain in approximately 40 percent of CBSAs. Payment adjustments of up to plus or minus nine percent apply to all Part B claims, not only episode-related services. CMS retains a portion of Part B reimbursement (1.35 percent rising to 1.80 percent) as direct Medicare savings. ASM runs from January 2027 through December 2031.
GLOBE (Global Benchmark for Efficient Drug Pricing), proposed December 2025, mandates international reference pricing for certain high-cost Part B drugs. Manufacturers must pay MFN-based rebates when U.S. prices exceed a benchmark derived from 19 OECD countries. The model covers approximately 25 percent of Part B beneficiaries in selected geographic regions and runs from October 2026 through September 2031.
GUARD (Guarding U.S. Medicare Against Rising Drug Costs), proposed December 2025, applies the same MFN methodology to Medicare Part D drugs. Mandatory for manufacturers of qualifying single-source drugs and biologicals, covering approximately 25 percent of Part D enrollees. January 2027 through December 2031.
Voluntary models with financial risk:
LEAD (Long-term Enhanced ACO Design), announced December 2025, succeeds ACO REACH as a voluntary 10-year total cost of care model for ACOs. It targets smaller, rural, and independent providers and emphasizes high-needs populations including dual eligibles and homebound beneficiaries. Two risk tracks: global (100 percent) and professional (50 percent). CMS Administered Risk Arrangements link ACOs with specialists through episode-based arrangements. January 2027 through December 2036.
BALANCE (Better Approaches to Lifestyle and Nutrition for Comprehensive hEalth), announced December 2025, creates Medicare Part D and Medicaid coverage for GLP-1 medications for weight management through CMS-negotiated manufacturer pricing. The $245 net price and $50 beneficiary copay structure represents the administration’s most politically visible CMMI model. Medicaid from May 2026, Medicare GLP-1 Bridge July through December 2026, full Part D from January 2027 through December 2031.
GENEROUS (GENErating cost Reductions fOr U.S. Medicaid), announced November 2025, provides MFN-level drug pricing for state Medicaid programs through voluntary manufacturer supplemental rebate agreements. Eight-country reference basket. January 2026 through December 2030.
Voluntary models without direct financial risk:
ACCESS (Advancing Chronic Care with Effective, Scalable Solutions), announced December 2025, creates outcome-aligned payments for technology-supported chronic disease management in Medicare FFS. Ten-year voluntary model open to Part B providers and digital health companies. Eligible conditions include hypertension, obesity, prediabetes, diabetes, CKD, cardiovascular disease, musculoskeletal pain, and depression/anxiety. July 2026 through June 2036.
MAHA ELEVATE (Make America Healthy Again: Enhancing Lifestyle and Evaluating Value-based Approaches Through Evidence), announced December 2025, funds up to 30 cooperative agreements at approximately $100 million total to test whole-person lifestyle medicine interventions not currently covered by Medicare. Three-year performance period, two cohorts starting September 2026 and 2027.
Major redesign of active model:
AHEAD (Achieving Healthcare Efficiency through Accountable Design), redesigned September 2025, extended all cohorts through December 2035, added Geo AHEAD geographic ACO program with non-provider entity eligibility and competitive bidding, added PC AHEAD capitated primary care payments, replaced Health Equity Plans with Population Health Accountability Plans, and imposed choice and competition policy requirements on participating states. Six active states with two new state slots opening July 2026. Geo AHEAD begins 2028.
The Portfolio Logic#
Several structural patterns emerge from the aggregate design.
The mandatory-voluntary split is condition-specific. Mandatory models target areas where CMS has identified specific waste or pricing disparity: unnecessary procedures (WISeR), specialist care for high-cost chronic conditions (ASM), and international drug price differentials (GLOBE, GUARD). Voluntary models target areas where CMS needs participation to achieve scale: ACO formation (LEAD), drug coverage expansion (BALANCE, GENEROUS), technology adoption (ACCESS), and lifestyle medicine evidence generation (MAHA ELEVATE). The administration is not uniformly mandatory or uniformly voluntary. It is mandatory where it believes it can document waste or overpayment, and voluntary where it needs willing participants to test new coverage or delivery models.
The chronic disease prevention thesis runs through everything. ACCESS addresses chronic conditions through technology. BALANCE provides pharmacological treatment for the obesity that drives chronic disease. MAHA ELEVATE tests behavioral interventions for chronic disease prevention. LEAD emphasizes evidence-based prevention within ACO care delivery. ASM holds specialists accountable for chronic disease management in heart failure and low back pain. AHEAD requires Population Health Accountability Plans focused on chronic disease prevention. The MAHA agenda is not a separate policy track layered on top of CMMI’s payment reform portfolio. It is embedded within the payment reform portfolio itself.
Drug pricing occupies a disproportionate share of the portfolio. Four of the ten models (BALANCE, GENEROUS, GLOBE, GUARD) address pharmaceutical pricing directly. A fifth (AHEAD) includes drug cost management within its total cost of care framework. The administration is using CMMI as its primary vehicle for drug pricing reform outside the IRA’s statutory negotiation process, creating a parallel pricing infrastructure that operates under demonstration authority rather than legislation.
Geographic concentration is a deliberate design choice. WISeR operates in six states. ASM in 40 percent of CBSAs. GLOBE and GUARD each in 25 percent of beneficiary regions. AHEAD in six states (expanding to eight). These geographic limitations serve dual purposes: they satisfy the statutory requirement that CMMI models be “tests” rather than national implementations, and they create natural control groups for evaluation. But they also mean that the lived experience of Medicare varies dramatically by geography. A beneficiary in a WISeR state, an ASM-selected CBSA, a GLOBE region, and an AHEAD state simultaneously inhabits a Medicare program that looks fundamentally different from the one experienced by a beneficiary in a state where none of these models operate.
What the Portfolio Does Not Include#
The absences are as revealing as the inclusions.
No MA-specific models. The 2025 portfolio contains no model that operates within Medicare Advantage. VBID was terminated. No replacement was announced. Every new model targets Medicare FFS, Medicaid, or pharmaceutical manufacturers. For a program that covers 54 percent of Medicare beneficiaries, MA’s absence from the Innovation Center’s 2025 agenda is conspicuous. The administration appears content to reshape MA through the rate-setting and risk adjustment process (the CY 2027 Advance Notice’s 0.09 percent increase and $7.2 billion chart review exclusion) rather than through CMMI model tests.
No home health model. The Home Health Value-Based Purchasing Model, the one CMMI model ever certified for national expansion, continues to operate. But the 2025 portfolio includes no new home health model, no model specifically targeting home-based care delivery, and no model testing home health payment innovation. AHEAD’s hospital global budgets create incentives to shift care out of hospitals (potentially into home settings), and LEAD’s emphasis on homebound beneficiaries creates demand for home-based providers. But neither model directly addresses home health payment or delivery.
No nursing facility or post-acute model. Skilled nursing facilities, inpatient rehabilitation, and long-term acute care hospitals are absent from the 2025 portfolio. The TEAM model (finalized in 2024 for a January 2026 start) covers post-acute care within surgical episode bundles, but no 2025 model specifically targets post-acute payment reform.
No beneficiary engagement or plan selection model. The 2025 portfolio tests payment to providers, pricing by manufacturers, and accountability for states. It does not test mechanisms for improving how beneficiaries choose among coverage options, navigate the system, or engage with their own care outside of MAHA ELEVATE’s small-scale lifestyle intervention grants.
Implementation Capacity as the Binding Constraint#
The aggregate implementation burden of the 2025 portfolio is staggering. CMS must simultaneously operate the WISeR prior authorization system across six states, finalize and implement the ASM mandatory specialist model, finalize and implement GLOBE and GUARD through rulemaking while defending likely litigation, negotiate and administer the BALANCE Bridge and the full BALANCE model across Medicaid and Part D, operate the GENEROUS manufacturer negotiation and state enrollment process, process LEAD applications and stand up a 10-year ACO model by January 2027, launch ACCESS with technology company applications and outcome-aligned payments, award and manage 30 MAHA ELEVATE cooperative agreements, and implement AHEAD’s redesigned framework across six states while onboarding two new states and building the Geo AHEAD competitive bidding infrastructure for 2028.
Each of these models requires dedicated CMS staff, contractor support, IT systems, rulemaking or application processing, monitoring infrastructure, and evaluation design. CMS is executing this portfolio while simultaneously managing the CY 2027 MA rate cycle, processing the Part D benefit redesign’s second year of implementation, conducting the third round of IRA drug price negotiations, and operating the existing MSSP, ACO REACH (through 2026), TEAM, and AHEAD programs.
The question is not whether any individual model is well-designed. Most are. The question is whether CMS has the institutional capacity to implement ten new models simultaneously while operating the existing Medicare program without implementation failures that undermine the models’ intended effects. The historical record is not encouraging: CMMI’s track record includes models that suffered from delayed launches, inadequate data infrastructure, participant confusion about requirements, and evaluation designs compromised by operational problems. Doing ten at once, across three programs, with four mandatory rulemaking processes, is an unprecedented test of CMS’s implementation bandwidth.
Where This Series Goes Next#
The ten articles in this series have traced the full arc of the 2025 CMMI reset: from the March cancellations through the May strategic refresh, the summer and fall model announcements, and the December portfolio completion. The models described here will generate their first performance data in 2026 and 2027. Their evaluations will take years. Their legal challenges will unfold concurrently. Their political sustainability will depend on whether the beneficiaries, providers, manufacturers, and states affected by them experience the models as improvements or impositions.
The remaining series in this publication examine how those experiences unfold: for MA plans navigating rate compression and risk adjustment reform (Series 2), for states managing the intersection of federal policy and local implementation (Series 3), for payers calculating whether MA remains viable (Series 4), for providers building strategy in a mandatory-risk environment (Series 5), for technology companies pursuing the ACCESS and AHEAD openings (Series 6), and for the 67 million Americans sitting at the kitchen table trying to understand what just changed about their Medicare (Series 7). The CMMI portfolio is the engine. Everything else is the road.
Related Reading#
MCR-00_01 The Trust Fund Clock MCR-03_05 CMS Under Pressure: Implementation Capacity, Workforce, and the Risk of Regulatory Overload MCR-12_01 The MA Plan Landscape Under Pressure: UnitedHealth, Humana, CVS/Aetna, Elevance, and the Regional Plans