The October 2025 SAT: the test where the numbers were letters

Yes, and unusually, students agree. October drew the sharpest difficulty reaction of any recent date, and it came from people who are normally fine. Tutors said the same thing about the second modules. If you walked out rattled, you had plenty of company.

Two complaints dominate. Reading and Writing Module 2 ate the clock, with dense passages rather than tricky questions. Math Module 2 turned abstract after the first third. If you got through Module 1 of either section thinking it was routine and then stalled, that was the shape of the day.

Worth saying plainly, because plenty of students left convinced they had wrecked their score. A form everybody finds hard is not a form you did badly on. Your score comes from what you did, not from how the room felt.

The items students remember from the back of Math Module 2 share one feature. They contain almost no numbers. You get constants and a condition, and you have to say what the condition forces.

One gave an equation built from letters and asked for one variable in terms of the others. There’s nothing to compute. You move every term containing your target to one side, factor it out, then divide. From 8rst equals 5rt plus 24s, you get s times 8rt minus 24 on the left, so s is 5rt over 8rt minus 24. The only way to go wrong is a sign, and signs are where most people lost it.

Another gave a quadratic with constants for coefficients and said it has two distinct real solutions. That’s a discriminant question wearing a disguise. For 2x squared plus bx plus 7c, the discriminant is b squared minus 56c, and two distinct solutions need it strictly positive. So b is greater than the square root of 56c. Use 28c and you forgot the leading 2. Allow equality and you let in the repeated root.

These questions test whether you can turn a condition into an equation, and nothing else. That’s why Desmos felt useless to so many students here. It graphs numbers. Give it a letter and you have to supply the relationship yourself before it can help at all.

Students describe the second reading module as long rather than tricky. Passages ran dense. One was a table of filing rules and penalties with four categories to keep straight.

That distinction changes what you should practice. If the questions were tricky, you would drill question types. They weren’t. The reading load was the problem, so drill the intake.

For a dense table, read the question first and find the one column it asks about. In the filing question, the claim concerned penalties on retained income, so only that column mattered. Reading down it gives none, 8 percent, 30 percent, none. The neighboring column, which carried 12 percent for everybody, is there to be read by accident. You can answer in fifteen seconds if you never look at it.

Geometry still led by volume of mentions, but the items that decided scores were algebraic. Circles, trigonometry and scale factors appeared alongside them. Each one came with a condition to convert rather than a value to compute.

The circle question is a good example. An equation in general form is tangent to a horizontal line, and you solve for the constant. Completing the square gives you a center and a radius squared of 17 minus c. Tangency means the vertical distance from the center to the line equals the radius, and that distance is 7, so 17 minus c equals 49 and c is negative 32. The word tangent did all the work.

Scale came up too, and the answers split exactly where you’d expect. Lengths in a drawing at one fifteenth of real size make areas one two-hundred-and-twenty-fifth. Students who answered one fifteenth used the length factor once. Volume would need the cube, so check which one the question wants before you square anything.

A parameter inequality caught people for a different reason. Given ax minus 18 is at least 12 for every x of at least 6, you need the smallest a that works. Rearrange to ax at least 30, note that negative a fails, then test the worst case, which is x equals 6. That gives a of at least 5. The question sounds like it could want a largest value, and reading it as a largest value leaves you with nothing to solve.

The reading section paired heavy passages with narrow questions. That’s a good trade if you know where to look, and a disaster if you read everything at the same speed.

Vocabulary showed the pattern most clearly. One question needed enumerate, and the clue sat in the next sentence, which reported a count of 214 references. The sentence with the blank in it was no help at all. Another needed speculative, and the deciding words were neither contradicts nor establishes. Both choices described reasonable scholarly attitudes. Only one matched evidence that settles nothing.

An inference question about how far people traveled to quarry a stone worked the same way. The finding was a falloff with distance. The conclusion was that people used the nearest source. That only follows if the sources were otherwise comparable, so the answer had to rule out differences in quality and access. The competing choice described a relationship between distance and yield, which sounds quantitative and was never established.

Conventions questions were the fast points. One tested each as a singular subject, where the verb later in the sentence already tells you the number. Another put a measurement in front of a complete second clause, so a comma alone spliced it and a semicolon did not.

Original questions written around what this exam tested, one for each argument above. Nothing here is a real test question.

October looks like a math problem and is mostly a translation problem. Tangent, two distinct solutions, holds for every x, infinitely many solutions. Each of those is an English sentence with exactly one equation behind it. Students who knew the equations finished those items fast. Students who went looking for numbers found none.

Build the list yourself. Take ten conditions that show up on real forms and write the equation each one forces, in your own words. It’s an hour of work, and it converts the part of the exam that felt impossible into the part you answer quickest.