grok-4.5

Grok 4.5 is SpaceXAI's frontier model built for coding, agentic tasks, and knowledge work. It was trained in SpaceXAI's data centers in Memphis with new datasets spanning science, engineering, and math.

Grok 4.5 isn't available in the API console for EU users yet. Availability is expected later this month.

If you already have an API key, set the model name to grok-4.5:

import os
from xai_sdk import Client
from xai_sdk.chat import user

client = Client(api_key=os.getenv("XAI_API_KEY"))

chat = client.chat.create(model="grok-4.5")
chat.append(user("Find and fix the bug, then explain it: function median(a){a.sort();return a[a.length/2]}"))

response = chat.sample()
print(response.content)

import { xai } from '@ai-sdk/xai';
import { generateText } from 'ai';

const { text } = await generateText({
  model: xai.responses('grok-4.5'),
  prompt:
    'Find and fix the bug, then explain it: function median(a){a.sort();return a[a.length/2]}',
});

console.log(text);

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.XAI_API_KEY,
  baseURL: 'https://api.x.ai/v1',
});

const response = await client.responses.create({
  model: 'grok-4.5',
  input: [
    {
      role: 'user',
      content:
        'Find and fix the bug, then explain it: function median(a){a.sort();return a[a.length/2]}',
    },
  ],
});

console.log(response.output_text);

curl https://api.x.ai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $XAI_API_KEY" \
  -d '{
    "model": "grok-4.5",
    "input": "Find and fix the bug, then explain it: function median(a){a.sort();return a[a.length/2]}"
  }'

New to the xAI API? Follow the Quickstart to create an account and make your first request.

Full context window, rate limits, and live pricing for your team are on the model detail page and Pricing.

For benchmark results and cost-versus-score comparisons, see the announcement.

• We highly recommend setting a prompt_cache_key (Responses API; x-grok-conv-id header on Chat Completions). It routes a conversation's requests to the same server, making cache hits reliable; without it you often pay full input price on a cache-cold server. See What Breaks Caching for common mistakes.
• Long agent loops additionally benefit from context compaction; for tool-heavy workloads see function calling.

• xAI API: get a key from the console
• Grok Build: the default model of the coding agent, on the API and CLI
• Cursor: available on all plans
• Office add-ins: default model in the Word, PowerPoint, and Excel add-ins
• Model gateways: OpenRouter, Vercel, Cloudflare, Snowflake, and Databricks Mosaic

• Reasoning - controlling reasoning_effort
• Announcement - launch post with demos and full benchmark figures
• Models - compare available models and their capabilities
• Pricing - token pricing for all models