As a fraction of GDP, that is.
The growth is mostly due to a) more recipients and b) each recipient receiving more in benefits, not more means-testing.
Here is the CBO report, and here is the relevant graph:
I should note that this graph is a little misleading, like all graphs that stack multiple components on top of one another, because if you look closely you will see that the increase is overwhelmingly driven by health care. And if you take a longer view, out to the 2020s, you will see (in the report) that means-tested health care spending (Medicaid, basically) will continue to rise as a fraction of GDP, whereas most other means-tested programs will continue to shrink.
In fact, means-tested cash assistance programs (TANF, SSI, EITC, & CTC), nutrition (SNAP, aka “food stamps”), education (Pell grants), and housing assistance have not grown much at all in the last 20 years — from about 1.5% of GDP in the early ’90s to about 2% now — and more importantly are scheduled to revert to the historical average within ten years, largely owing to slow but steady economic recovery.
Bottom line: non-health means-tested programs have certainly not been gutted (contra this for example), nor have they been very stingy, but at the same time let no one say that they have “exploded” or that they represent some dire fiscal threat. They are reasonably cheap (1 – 2% of GDP) yet effective programs, disability insurance notwithstanding. Health care, however, both means-tested and otherwise, is still the biggest unresolved social and fiscal challenge.