diff --git a/src/xy/__tests__/xyBinning.test.ts b/src/xy/__tests__/xyBinning.test.ts
new file mode 100644
index 00000000..6e518413
--- /dev/null
+++ b/src/xy/__tests__/xyBinning.test.ts
@@ -0,0 +1,243 @@
+import { expect, test } from 'vitest';
+
+import { xyBinning } from '../xyBinning.ts';
+
+test('binSize 1 returns a copy', () => {
+  const data = {
+    x: [1, 2, 3, 4],
+    y: [10, 20, 30, 40],
+  };
+
+  expect(xyBinning(data, { binSize: 1 })).toStrictEqual({
+    x: Float64Array.from([1, 2, 3, 4]),
+    y: Float64Array.from([10, 20, 30, 40]),
+  });
+});
+
+test('default keepFirstAndLast with even division', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6],
+    y: [10, 20, 30, 40, 50, 60],
+  };
+
+  expect(xyBinning(data, { binSize: 2 })).toStrictEqual({
+    x: Float64Array.from([1, 2.5, 4.5, 6]),
+    y: Float64Array.from([10, 25, 45, 60]),
+  });
+  expect(xyBinning(data, { binSize: 3 })).toStrictEqual({
+    x: Float64Array.from([1, 3, 5, 6]),
+    y: Float64Array.from([10, 30, 50, 60]),
+  });
+});
+
+test('default keepFirstAndLast with uneven division', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6, 7],
+    y: [10, 20, 30, 40, 50, 60, 70],
+  };
+
+  expect(xyBinning(data, { binSize: 3 })).toStrictEqual({
+    x: Float64Array.from([1, 3, 5.5, 7]),
+    y: Float64Array.from([10, 30, 55, 70]),
+  });
+});
+
+test('accepts Float64Array input', () => {
+  const data = {
+    x: Float64Array.from([2, 4, 6, 8]),
+    y: Float64Array.from([20, 40, 60, 80]),
+  };
+
+  expect(xyBinning(data, { binSize: 2 })).toStrictEqual({
+    x: Float64Array.from([2, 5, 8]),
+    y: Float64Array.from([20, 50, 80]),
+  });
+});
+
+test('x is centered in each bin', () => {
+  const data = {
+    x: [1, 2, 10, 20],
+    y: [10, 20, 30, 40],
+  };
+
+  expect(
+    xyBinning(data, { binSize: 2, keepFirstAndLast: false }),
+  ).toStrictEqual({
+    x: Float64Array.from([1.5, 15]),
+    y: Float64Array.from([15, 35]),
+  });
+});
+
+test('default binSize is 10', () => {
+  const data = {
+    x: Array.from({ length: 25 }, (_, i) => i + 1),
+    y: Array.from({ length: 25 }, (_, i) => (i + 1) * 10),
+  };
+
+  expect(xyBinning(data)).toStrictEqual({
+    x: Float64Array.from([1, 6.5, 16.5, 23, 25]),
+    y: Float64Array.from([10, 65, 165, 230, 250]),
+  });
+});
+
+test('keepFirstAndLast=false restores pure binning', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6],
+    y: [10, 20, 30, 40, 50, 60],
+  };
+
+  expect(
+    xyBinning(data, { binSize: 2, keepFirstAndLast: false }),
+  ).toStrictEqual({
+    x: Float64Array.from([1.5, 3.5, 5.5]),
+    y: Float64Array.from([15, 35, 55]),
+  });
+  expect(
+    xyBinning(data, { binSize: 3, keepFirstAndLast: false }),
+  ).toStrictEqual({
+    x: Float64Array.from([2, 5]),
+    y: Float64Array.from([20, 50]),
+  });
+});
+
+test('keepFirstAndLast=false with uneven division', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6, 7],
+    y: [10, 20, 30, 40, 50, 60, 70],
+  };
+
+  expect(
+    xyBinning(data, { binSize: 3, keepFirstAndLast: false }),
+  ).toStrictEqual({
+    x: Float64Array.from([2, 5, 7]),
+    y: Float64Array.from([20, 50, 70]),
+  });
+});
+
+test('throws on invalid binSize', () => {
+  const data = {
+    x: [1, 2, 3],
+    y: [10, 20, 30],
+  };
+
+  expect(() => xyBinning(data, { binSize: 0 })).toThrow(
+    /binSize must be a positive integer/,
+  );
+  expect(() => xyBinning(data, { binSize: 1.5 })).toThrow(
+    /binSize must be a positive integer/,
+  );
+  expect(() => xyBinning(data, { binSize: -2 })).toThrow(
+    /binSize must be a positive integer/,
+  );
+});
+
+test('throws on empty input', () => {
+  expect(() => xyBinning({ x: [], y: [] }, { binSize: 2 })).toThrow(
+    /input must not be empty/,
+  );
+});
+
+test('throws when x does not contain numbers', () => {
+  const data = {
+    x: ['a', 'b', 'c'],
+    y: [1, 2, 3],
+  };
+
+  expect(() => xyBinning(data as any, { binSize: 2 })).toThrow(
+    /input must contain numbers/,
+  );
+});
+
+test('numberOfPoints splits into N bins', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
+    y: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
+  };
+
+  expect(
+    xyBinning(data, { numberOfPoints: 5, keepFirstAndLast: false }),
+  ).toStrictEqual({
+    x: Float64Array.from([1.5, 3.5, 5.5, 7.5, 9.5]),
+    y: Float64Array.from([1.5, 3.5, 5.5, 7.5, 9.5]),
+  });
+});
+
+test('numberOfPoints with uneven split', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
+    y: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
+  };
+
+  expect(
+    xyBinning(data, { numberOfPoints: 3, keepFirstAndLast: false }),
+  ).toStrictEqual({
+    x: Float64Array.from([2, 5, 8.5]),
+    y: Float64Array.from([2, 5, 8.5]),
+  });
+});
+
+test('numberOfPoints with keepFirstAndLast', () => {
+  const data = {
+    x: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
+    y: [10, 20, 30, 40, 50, 60, 70, 80, 90, 100],
+  };
+
+  expect(xyBinning(data, { numberOfPoints: 5 })).toStrictEqual({
+    x: Float64Array.from([1, 2.5, 5, 8, 10]),
+    y: Float64Array.from([10, 25, 50, 80, 100]),
+  });
+});
+
+test('numberOfPoints throws when > length', () => {
+  expect(() =>
+    xyBinning({ x: [1, 2, 3], y: [10, 20, 30] }, { numberOfPoints: 4 }),
+  ).toThrow(/numberOfPoints must be <= array.length/);
+});
+
+test('numberOfPoints throws when not a positive integer', () => {
+  expect(() =>
+    xyBinning({ x: [1, 2, 3], y: [10, 20, 30] }, { numberOfPoints: 0 }),
+  ).toThrow(/numberOfPoints must be a positive integer/);
+  expect(() =>
+    xyBinning({ x: [1, 2, 3], y: [10, 20, 30] }, { numberOfPoints: 2.5 }),
+  ).toThrow(/numberOfPoints must be a positive integer/);
+});
+
+test('binSize and numberOfPoints are mutually exclusive', () => {
+  expect(() =>
+    xyBinning(
+      { x: [1, 2, 3, 4], y: [10, 20, 30, 40] },
+      { binSize: 2, numberOfPoints: 2 },
+    ),
+  ).toThrow(/mutually exclusive/);
+});
+
+test('numberOfPoints=2 with keepFirstAndLast returns endpoints', () => {
+  expect(
+    xyBinning(
+      { x: [1, 2, 3, 4, 5], y: [10, 20, 30, 40, 50] },
+      { numberOfPoints: 2 },
+    ),
+  ).toStrictEqual({
+    x: Float64Array.from([1, 5]),
+    y: Float64Array.from([10, 50]),
+  });
+});
+
+test('numberOfPoints < 2 with keepFirstAndLast throws', () => {
+  expect(() =>
+    xyBinning({ x: [1, 2, 3], y: [10, 20, 30] }, { numberOfPoints: 1 }),
+  ).toThrow(/numberOfPoints must be >= 2 when keepFirstAndLast is true/);
+});
+
+test('numberOfPoints=1 with keepFirstAndLast=false averages entire array', () => {
+  expect(
+    xyBinning(
+      { x: [1, 2, 3, 4, 5], y: [10, 20, 30, 40, 50] },
+      { numberOfPoints: 1, keepFirstAndLast: false },
+    ),
+  ).toStrictEqual({
+    x: Float64Array.from([3]),
+    y: Float64Array.from([30]),
+  });
+});
diff --git a/src/xy/xyBinning.ts b/src/xy/xyBinning.ts
new file mode 100644
index 00000000..ec7b3733
--- /dev/null
+++ b/src/xy/xyBinning.ts
@@ -0,0 +1,25 @@
+import type { DataXY } from 'cheminfo-types';
+
+import type { XBinningOptions } from '../x/index.ts';
+import { xBinning } from '../x/index.ts';
+
+import { xyCheck } from './xyCheck.ts';
+
+/**
+ * Downsample x and y by averaging consecutive non-overlapping points.
+ * Both arrays are binned with the same partitioning, so the resulting x
+ * is the mean x of each bin.
+ * @param data - Object that contains x and y arrays.
+ * @param options - Binning options (same as xBinning).
+ * @returns Downsampled x and y arrays.
+ */
+export function xyBinning(
+  data: DataXY,
+  options: XBinningOptions = {},
+): DataXY<Float64Array> {
+  xyCheck(data);
+  return {
+    x: xBinning(data.x, options),
+    y: xBinning(data.y, options),
+  };
+}