diff --git a/README.md b/README.md
index 55c7326..9edfe7d 100644
--- a/README.md
+++ b/README.md
@@ -125,6 +125,7 @@ on the target page.
 | GeeTest v4 | ✅ | ✅ |
 | Image to Text | ✅ | ❌ |
 | Yandex SmartCaptcha | ✅ | ✅ |
+| Yandex SmartCaptcha Image | ✅ | ❌ |
 | Coordinates (click captcha) | ✅ | ❌ |
 | Tencent | ✅ | ✅ |
 
@@ -450,10 +451,10 @@ With proxy, use `Tasks.GeeTest` (same proxy fields as reCAPTCHA v2).
 <sup>[API method description.](https://captcha-solver.com/en/docs/captcha-types#yandex-smartcaptcha)</sup>
 
 Use this method to solve the token-based Yandex SmartCaptcha and obtain a token
-for the widget on the target page. Use the coordinates method for image challenges.
+for the widget on the target page. Use the dedicated image method for image challenges.
 
 `Tasks.YandexSmartCaptchaTaskProxyless` / `Tasks.YandexSmartCaptchaTask` -- token-based
-challenge. For the image challenge instead, use `Tasks.CoordinatesTask` with
+challenge.
 
 | Field | Required | Description |
 |---|---|---|
@@ -481,15 +482,41 @@ console.log(result.token);
 With proxy, use `Tasks.YandexSmartCaptchaTask` (same proxy fields as reCAPTCHA v2,
 plus the `https` option above).
 
+#### Image challenge
+
+Use `Tasks.YandexSmartCaptchaImageTask` when SmartCaptcha shows a main image and
+a separate strip indicating the symbols to click. Encode both images as base64;
+the SDK sends API v2 type `SmartCaptchaTask`, and the returned coordinates are
+already ordered for clicking.
+
+| Field | Required | Description |
+|---|---|---|
+| `image` | yes | Main captcha image, base64-encoded. |
+| `imgInstructions` | yes | Click-order instruction image, base64-encoded. |
+| `comment` | no | Optional text hint for the worker. |
+
+```javascript
+import fs from 'node:fs';
+import { CaptchaClient, Tasks } from '@captcha-solver-api/javascript-sdk';
+
+const task = new Tasks.YandexSmartCaptchaImageTask({
+  image: fs.readFileSync('captcha.png').toString('base64'),
+  imgInstructions: fs.readFileSync('instructions.png').toString('base64')
+});
+const result = await new CaptchaClient({ clientKey: 'your_api_key' }).solve(task);
+console.log(result.coordinates);
+```
+
+See the runnable [async](examples/async/yandex_smartcaptcha_image.js) and
+[promise-chain](examples/sync/yandex_smartcaptcha_image.js) examples.
+
 ### Coordinates (click captcha)
 
 <sup>[API method description.](https://captcha-solver.com/en/docs/captcha-types#coordinates)</sup>
 
-Use this method to identify points that a worker should click in an image. It
-supports generic click captchas and the image version of Yandex SmartCaptcha.
+Use this method to identify points that a worker should click in a generic image.
 
-`Tasks.CoordinatesTask`. Used both for generic "click on X" captchas and for
-directly.
+`Tasks.CoordinatesTask`.
 
 | Field | Required | Description |
 |---|---|---|
diff --git a/examples/README.md b/examples/README.md
index 044fbd6..afbb024 100644
--- a/examples/README.md
+++ b/examples/README.md
@@ -78,6 +78,7 @@ twice the balance.
 | GeeTest v3 | [geetest_v3.js](async/geetest_v3.js) | [geetest_v3.js](sync/geetest_v3.js) | `GeeTestProxyless`, `GeeTest` | `challenge`, `validate`, `seccode` |
 | GeeTest v4 | [geetest_v4.js](async/geetest_v4.js) | [geetest_v4.js](sync/geetest_v4.js) | `GeeTestProxyless`, `GeeTest` | `captcha_id`, `lot_number`, `pass_token`, `gen_time`, `captcha_output` |
 | Yandex SmartCaptcha | [yandex_smartcaptcha.js](async/yandex_smartcaptcha.js) | [yandex_smartcaptcha.js](sync/yandex_smartcaptcha.js) | `YandexSmartCaptchaTaskProxyless`, `YandexSmartCaptchaTask` | `token` |
+| Yandex SmartCaptcha Image | [yandex_smartcaptcha_image.js](async/yandex_smartcaptcha_image.js) | [yandex_smartcaptcha_image.js](sync/yandex_smartcaptcha_image.js) | `YandexSmartCaptchaImageTask` | `coordinates` |
 | Tencent | [tencent.js](async/tencent.js) | [tencent.js](sync/tencent.js) | `TencentTaskProxyless`, `TencentTask` | `appid`, `ret`, `ticket`, `randstr` |
 | Image to Text | [image_to_text.js](async/image_to_text.js) | [image_to_text.js](sync/image_to_text.js) | `ImageToText` | `text` |
 | Coordinates (click) | [coordinates.js](async/coordinates.js) | [coordinates.js](sync/coordinates.js) | `CoordinatesTask` | `coordinates` |
diff --git a/examples/assets/yandex-smartcaptcha-image.png b/examples/assets/yandex-smartcaptcha-image.png
new file mode 100644
index 0000000..92addc3
Binary files /dev/null and b/examples/assets/yandex-smartcaptcha-image.png differ
diff --git a/examples/assets/yandex-smartcaptcha-instructions.png b/examples/assets/yandex-smartcaptcha-instructions.png
new file mode 100644
index 0000000..80cf090
Binary files /dev/null and b/examples/assets/yandex-smartcaptcha-instructions.png differ
diff --git a/examples/async/README.md b/examples/async/README.md
index 9015c41..189bff8 100644
--- a/examples/async/README.md
+++ b/examples/async/README.md
@@ -28,6 +28,7 @@ spends balance twice. Comment out the one you don't need.
 | [geetest_v3.js](geetest_v3.js) | Fetches a fresh `challenge`, then solves GeeTest v3 | Fetch · Proxyless · With proxy |
 | [geetest_v4.js](geetest_v4.js) | Solves GeeTest v4 via `captcha_id` | Proxyless · With proxy |
 | [yandex_smartcaptcha.js](yandex_smartcaptcha.js) | Solves Yandex SmartCaptcha (token challenge) | Proxyless · With proxy |
+| [yandex_smartcaptcha_image.js](yandex_smartcaptcha_image.js) | Solves Yandex SmartCaptcha image challenge | single task |
 | [tencent.js](tencent.js) | Solves a Tencent captcha | Proxyless · With proxy |
 | [image_to_text.js](image_to_text.js) | Recognises text on a captcha image | Basic · Advanced · With language pool |
 
diff --git a/examples/async/yandex_smartcaptcha_image.js b/examples/async/yandex_smartcaptcha_image.js
new file mode 100644
index 0000000..e878647
--- /dev/null
+++ b/examples/async/yandex_smartcaptcha_image.js
@@ -0,0 +1,26 @@
+/** Solve the image-based Yandex SmartCaptcha challenge. */
+import 'dotenv/config';
+import fs from 'node:fs';
+import { CaptchaClient, Tasks } from '@captcha-solver-api/javascript-sdk';
+
+const captchaSolver = new CaptchaClient({
+  clientKey: process.env.CAPTCHA_API_KEY || 'YOUR_API_KEY'
+});
+
+const image = fs.readFileSync(
+  new URL('../assets/yandex-smartcaptcha-image.png', import.meta.url)
+).toString('base64');
+const imgInstructions = fs.readFileSync(
+  new URL('../assets/yandex-smartcaptcha-instructions.png', import.meta.url)
+).toString('base64');
+
+try {
+  const task = new Tasks.YandexSmartCaptchaImageTask({
+    image,
+    imgInstructions
+  });
+  const result = await captchaSolver.solve(task);
+  console.log('Coordinates in click order:', result.coordinates);
+} catch (error) {
+  console.error(error);
+}
diff --git a/examples/sync/README.md b/examples/sync/README.md
index 88fd009..30eaad4 100644
--- a/examples/sync/README.md
+++ b/examples/sync/README.md
@@ -35,6 +35,7 @@ Errors are handled per chain with a trailing `.catch()`; a failing block does no
 | [geetest_v3.js](geetest_v3.js) | Fetches a fresh `challenge`, then solves GeeTest v3 | one nested chain |
 | [geetest_v4.js](geetest_v4.js) | Solves GeeTest v4 via `captcha_id` | Proxyless · With proxy |
 | [yandex_smartcaptcha.js](yandex_smartcaptcha.js) | Solves Yandex SmartCaptcha (token challenge) | Proxyless · With proxy |
+| [yandex_smartcaptcha_image.js](yandex_smartcaptcha_image.js) | Solves Yandex SmartCaptcha image challenge | single chain |
 | [tencent.js](tencent.js) | Solves a Tencent captcha | Proxyless · With proxy |
 | [image_to_text.js](image_to_text.js) | Recognises text on a captcha image | Basic · Advanced · With language pool |
 
@@ -154,10 +155,7 @@ directory** and the script runs as-is. Solution: `text`.
 Three chains, no proxy variant, all active. *Basic* (`basicTask`) submits the shipped grid captcha
 (`coordinates-captcha.png`) plus a `comment` telling the worker what to click. *Advanced*
 (`advancedTask`) solves a different image, a real traffic-lights photo (`traffic-lights.png`), with
-`traffic-lights-instructions.png` as `imgInstructions` and `minClicks`/`maxClicks` limits. *Yandex
-SmartCaptcha image mode* (`yandexTask`) reuses the same `CoordinatesTask` with
-click-order hint is baked into the image itself, so no separate `imgInstructions` is passed — this is
-how the image variant of Yandex SmartCaptcha is solved.
+`traffic-lights-instructions.png` as `imgInstructions` and `minClicks`/`maxClicks` limits.
 
 Reads its images via
 `new URL(..., import.meta.url)`, so the path holds **regardless of the working directory** and the
diff --git a/examples/sync/yandex_smartcaptcha_image.js b/examples/sync/yandex_smartcaptcha_image.js
new file mode 100644
index 0000000..e56f4c3
--- /dev/null
+++ b/examples/sync/yandex_smartcaptcha_image.js
@@ -0,0 +1,28 @@
+/** Solve the image-based Yandex SmartCaptcha challenge using promises. */
+import 'dotenv/config';
+import fs from 'node:fs';
+import { CaptchaClient, Tasks } from '@captcha-solver-api/javascript-sdk';
+
+const captchaSolver = new CaptchaClient({
+  clientKey: process.env.CAPTCHA_API_KEY || 'YOUR_API_KEY'
+});
+
+const image = fs.readFileSync(
+  new URL('../assets/yandex-smartcaptcha-image.png', import.meta.url)
+).toString('base64');
+const imgInstructions = fs.readFileSync(
+  new URL('../assets/yandex-smartcaptcha-instructions.png', import.meta.url)
+).toString('base64');
+
+const task = new Tasks.YandexSmartCaptchaImageTask({
+  image,
+  imgInstructions
+});
+
+captchaSolver.solve(task)
+  .then((result) => {
+    console.log('Coordinates in click order:', result.coordinates);
+  })
+  .catch((error) => {
+    console.error(error);
+  });
diff --git a/src/index.ts b/src/index.ts
index 5620434..320b475 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -30,6 +30,7 @@ export type {
   YandexSmartCaptchaTaskProxylessParams,
   YandexSmartCaptchaTaskParams,
   CoordinatesTaskParams,
+  YandexSmartCaptchaImageTaskParams,
   TencentTaskProxylessParams,
   TencentTaskParams
 } from './tasks.js';
diff --git a/src/tasks.ts b/src/tasks.ts
index 7851c95..c27d615 100644
--- a/src/tasks.ts
+++ b/src/tasks.ts
@@ -553,7 +553,6 @@ export interface CoordinatesTaskParams {
 
 /**
  * Click captcha: the worker marks points in an image (`CoordinatesTask`).
- * Used for generic "click on X" captchas and the image version of Yandex SmartCaptcha.
  * Solution: {@link CoordinatesSolution}.
  *
  * @see https://captcha-solver.com/en/docs/captcha-types#coordinates
@@ -564,6 +563,35 @@ export class CoordinatesTask extends BaseTask<CoordinatesSolution> {
   }
 }
 
+/* ------------------------------------------------------------------ */
+/* Yandex SmartCaptcha image                                          */
+/* ------------------------------------------------------------------ */
+
+/** Parameters of {@link YandexSmartCaptchaImageTask}. */
+export interface YandexSmartCaptchaImageTaskParams {
+  /** The main SmartCaptcha challenge image, base64-encoded. */
+  image: string;
+  /** The SmartCaptcha click-order instruction image, base64-encoded. */
+  imgInstructions: string;
+  /** Optional text hint shown to the worker. */
+  comment?: string | null;
+}
+
+/**
+ * Image-based Yandex SmartCaptcha challenge.
+ *
+ * Serialized as the API v2 `SmartCaptchaTask`. Coordinates are returned in
+ * the order in which the symbols must be clicked.
+ * Solution: {@link CoordinatesSolution}.
+ *
+ * @see https://captcha-solver.com/en/docs/captcha-types#yandex-smartcaptcha
+ */
+export class YandexSmartCaptchaImageTask extends BaseTask<CoordinatesSolution> {
+  constructor({ image, imgInstructions, comment = null }: YandexSmartCaptchaImageTaskParams) {
+    super({ type: 'SmartCaptchaTask', image, imgInstructions, comment });
+  }
+}
+
 /* ------------------------------------------------------------------ */
 /* Tencent                                                             */
 /* ------------------------------------------------------------------ */
diff --git a/tests/integration/coordinates.test.js b/tests/integration/coordinates.test.js
index 828e62b..21ec678 100644
--- a/tests/integration/coordinates.test.js
+++ b/tests/integration/coordinates.test.js
@@ -20,13 +20,21 @@ import { Tasks } from '@captcha-solver-api/javascript-sdk';
 import { describeIntegration, createClient } from './helpers.js';
 
 const imagePath = new URL('../../examples/assets/coordinates-captcha.png', import.meta.url);
+const yandexImagePath = new URL(
+  '../../examples/assets/yandex-smartcaptcha-image.png',
+  import.meta.url
+);
+const yandexInstructionsPath = new URL(
+  '../../examples/assets/yandex-smartcaptcha-instructions.png',
+  import.meta.url
+);
 
 // The sample is a 4x4 grid captcha whose own header reads "Select all squares
 // with street signs". The comment repeats that instruction for the worker,
 // because nothing guarantees they read the text baked into the image.
 const comment = 'click on all squares with street signs';
 
-describeIntegration('Coordinates against the real API', () => {
+describeIntegration('Image coordinate tasks against the real API', () => {
   test('solve returns the clicked points', async () => {
     const image = fs.readFileSync(imagePath);
 
@@ -58,4 +66,28 @@ describeIntegration('Coordinates against the real API', () => {
     }
   }, 130000);
 
+  test('Yandex SmartCaptcha image returns ordered click coordinates', async () => {
+    const image = fs.readFileSync(yandexImagePath);
+    const instructions = fs.readFileSync(yandexInstructionsPath);
+    const task = new Tasks.YandexSmartCaptchaImageTask({
+      image: image.toString('base64'),
+      imgInstructions: instructions.toString('base64')
+    });
+
+    const solution = await createClient().solve(task);
+
+    expect(solution.coordinates).toHaveLength(5);
+
+    const width = image.readUInt32BE(16);
+    const height = image.readUInt32BE(20);
+    for (const point of solution.coordinates) {
+      expect(typeof point.x).toBe('number');
+      expect(typeof point.y).toBe('number');
+      expect(point.x).toBeGreaterThanOrEqual(0);
+      expect(point.y).toBeGreaterThanOrEqual(0);
+      expect(point.x).toBeLessThanOrEqual(width);
+      expect(point.y).toBeLessThanOrEqual(height);
+    }
+  }, 130000);
+
 });
diff --git a/tests/unit/public-api.test.js b/tests/unit/public-api.test.js
index 8b62af7..3f6066f 100644
--- a/tests/unit/public-api.test.js
+++ b/tests/unit/public-api.test.js
@@ -30,6 +30,7 @@ const TASK_CLASSES = [
   'YandexSmartCaptchaTaskProxyless',
   'YandexSmartCaptchaTask',
   'CoordinatesTask',
+  'YandexSmartCaptchaImageTask',
   'TencentTaskProxyless',
   'TencentTask'
 ];
diff --git a/tests/unit/yandex_smartcaptcha_image.test.js b/tests/unit/yandex_smartcaptcha_image.test.js
new file mode 100644
index 0000000..e9705cf
--- /dev/null
+++ b/tests/unit/yandex_smartcaptcha_image.test.js
@@ -0,0 +1,18 @@
+import { Tasks } from '@captcha-solver-api/javascript-sdk';
+
+describe('YandexSmartCaptchaImageTask', () => {
+  test('serializes as an API v2 SmartCaptchaTask with both images', () => {
+    const task = new Tasks.YandexSmartCaptchaImageTask({
+      image: 'captcha-image',
+      imgInstructions: 'instruction-image',
+      comment: 'click the symbols in the specified order'
+    });
+
+    expect(task.toDict()).toEqual({
+      type: 'SmartCaptchaTask',
+      image: 'captcha-image',
+      imgInstructions: 'instruction-image',
+      comment: 'click the symbols in the specified order'
+    });
+  });
+});