diff --git a/README.md b/README.md
index 4583e78..a05f962 100644
--- a/README.md
+++ b/README.md
@@ -186,10 +186,10 @@ Alerts are delivered as rich embeds with color-coded sidebars: red for FLASH, ye
 **Optional dependency:** The full bot requires `discord.js`. Install it with `npm install discord.js`. If it's not installed, Crucix automatically falls back to webhook-only mode.
 
 ### Optional LLM Layer
-Connect any of 6 LLM providers for enhanced analysis:
+Connect any of 8 LLM providers for enhanced analysis:
 - **AI trade ideas** — quantitative analyst producing 5-8 actionable ideas citing specific data
 - **Smarter alert evaluation** — LLM classifies signals into FLASH/PRIORITY/ROUTINE tiers with cross-domain correlation and confidence scoring
-- Providers: Anthropic Claude, OpenAI, Google Gemini, OpenRouter (Unified API), OpenAI Codex (ChatGPT subscription), MiniMax, Mistral
+- Providers: Anthropic Claude, OpenAI, Google Gemini, OpenRouter (Unified API), OpenAI Codex (ChatGPT subscription), MiniMax, Mistral, Grok
 - Graceful fallback — when LLM is unavailable, a rule-based engine takes over alert evaluation. LLM failures never crash the sweep cycle.
 
 ---
@@ -303,12 +303,12 @@ crucix/
 │       └── jarvis.html        # Self-contained Jarvis HUD
 │
 ├── lib/
-│   ├── llm/                   # LLM abstraction (5 providers, raw fetch, no SDKs)
+│   ├── llm/                   # LLM abstraction (8 providers, raw fetch, no SDKs)
 │   │   ├── provider.mjs       # Base class
 │   │   ├── anthropic.mjs      # Claude
 │   │   ├── openai.mjs         # GPT
 │   │   ├── gemini.mjs         # Gemini
-├──     ├── grok.mjs           # Grok
+│   │   ├── grok.mjs           # Grok
 │   │   ├── openrouter.mjs     # OpenRouter (Unified API)
 │   │   ├── codex.mjs          # Codex (ChatGPT subscription)
 │   │   ├── minimax.mjs        # MiniMax (M2.5, 204K context)
@@ -414,7 +414,7 @@ All settings are in `.env` with sensible defaults:
 |----------|---------|-------------|
 | `PORT` | `3117` | Dashboard server port |
 | `REFRESH_INTERVAL_MINUTES` | `15` | Auto-refresh interval |
-| `LLM_PROVIDER` | disabled | `anthropic`, `openai`, `gemini`, `codex`, `openrouter`, `minimax`, or `mistral` |
+| `LLM_PROVIDER` | disabled | `anthropic`, `openai`, `gemini`, `codex`, `openrouter`, `minimax`, `mistral`, or `grok` |
 | `LLM_API_KEY` | — | API key (not needed for codex) |
 | `LLM_MODEL` | per-provider default | Override model selection |
 | `TELEGRAM_BOT_TOKEN` | disabled | For Telegram alerts + bot commands |
diff --git a/lib/llm/grok.mjs b/lib/llm/grok.mjs
index 621bb7f..76fef66 100644
--- a/lib/llm/grok.mjs
+++ b/lib/llm/grok.mjs
@@ -22,7 +22,7 @@ export class GrokProvider extends LLMProvider {
         Authorization: `Bearer ${this.apiKey}`,
       },
       body: JSON.stringify({
-        max_tokens: opts.max_tokens || 4096,
+        max_tokens: opts.maxTokens || 4096,
         messages: [
           { role: 'system', content: systemPrompt },
           { role: 'user', content: userMessage },
diff --git a/test/llm-grok.test.mjs b/test/llm-grok.test.mjs
index 59a648a..1d400fe 100644
--- a/test/llm-grok.test.mjs
+++ b/test/llm-grok.test.mjs
@@ -1,7 +1,7 @@
 // Grok provider — unit tests
 // Uses Node.js built-in test runner (node:test) — no extra dependencies
 
-import { describe, it, mock, beforeEach } from 'node:test';
+import { describe, it, mock } from 'node:test';
 import assert from 'node:assert/strict';
 import { GrokProvider } from '../lib/llm/grok.mjs';
 import { createLLMProvider } from '../lib/llm/index.mjs';
@@ -9,12 +9,12 @@ import { createLLMProvider } from '../lib/llm/index.mjs';
 // ─── Unit Tests ───
 
 describe('GrokProvider', () => {
-  it('should set defaults correctly', () => {
-    const provider = new GrokProvider({ apiKey: 'sk-test' });
-    assert.equal(provider.name, 'grok');
-    assert.equal(provider.model, 'grok-3');
-    assert.equal(provider.isConfigured, true);
-  });
+  it('should set defaults correctly', () => {
+    const provider = new GrokProvider({ apiKey: 'sk-test' });
+    assert.equal(provider.name, 'grok');
+    assert.equal(provider.model, 'grok-4-latest');
+    assert.equal(provider.isConfigured, true);
+  });
 
   it('should accept custom model', () => {
     const provider = new GrokProvider({ apiKey: 'sk-test', model: 'grok-2' });
@@ -45,11 +45,11 @@ describe('GrokProvider', () => {
     }
   });
 
-  it('should parse successful response', async () => {
-    const provider = new GrokProvider({ apiKey: 'sk-test' });
-    const mockResponse = {
-      choices: [{ message: { content: 'Hello world' } }],
-      usage: { prompt_tokens: 10, completion_tokens: 5 },
+  it('should parse successful response', async () => {
+    const provider = new GrokProvider({ apiKey: 'sk-test' });
+    const mockResponse = {
+      choices: [{ message: { content: 'Hello world' } }],
+      usage: { prompt_tokens: 10, completion_tokens: 5 },
       model: 'grok-3'
     };
     const originalFetch = globalThis.fetch;
@@ -63,9 +63,63 @@ describe('GrokProvider', () => {
       assert.equal(result.usage.outputTokens, 5);
       assert.equal(result.model, 'grok-3');
     } finally {
-      globalThis.fetch = originalFetch;
-    }
-  });
+      globalThis.fetch = originalFetch;
+    }
+  });
+
+  it('should send correct request format', async () => {
+    const provider = new GrokProvider({ apiKey: 'sk-test-key', model: 'grok-4-latest' });
+    let capturedUrl, capturedOpts;
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn((url, opts) => {
+      capturedUrl = url;
+      capturedOpts = opts;
+      return Promise.resolve({
+        ok: true,
+        json: () => Promise.resolve({
+          choices: [{ message: { content: 'ok' } }],
+          usage: { prompt_tokens: 1, completion_tokens: 1 },
+          model: 'grok-4-latest',
+        }),
+      });
+    });
+    try {
+      await provider.complete('system prompt', 'user message', { maxTokens: 2048 });
+      assert.equal(capturedUrl, 'https://api.x.ai/v1/chat/completions');
+      assert.equal(capturedOpts.method, 'POST');
+      const headers = capturedOpts.headers;
+      assert.equal(headers['Content-Type'], 'application/json');
+      assert.equal(headers.Authorization, 'Bearer sk-test-key');
+      const body = JSON.parse(capturedOpts.body);
+      assert.equal(body.model, 'grok-4-latest');
+      assert.equal(body.max_tokens, 2048);
+      assert.equal(body.messages[0].role, 'system');
+      assert.equal(body.messages[0].content, 'system prompt');
+      assert.equal(body.messages[1].role, 'user');
+      assert.equal(body.messages[1].content, 'user message');
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
+
+  it('should handle empty response gracefully', async () => {
+    const provider = new GrokProvider({ apiKey: 'sk-test' });
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn(() =>
+      Promise.resolve({
+        ok: true,
+        json: () => Promise.resolve({ choices: [], usage: {} }),
+      })
+    );
+    try {
+      const result = await provider.complete('sys', 'user');
+      assert.equal(result.text, '');
+      assert.equal(result.usage.inputTokens, 0);
+      assert.equal(result.usage.outputTokens, 0);
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
 });
 
 // ─── Factory Tests ───
@@ -76,4 +130,4 @@ describe('createLLMProvider', () => {
     assert.ok(provider instanceof GrokProvider);
     assert.equal(provider.isConfigured, true);
   });
-});
\ No newline at end of file
+});