xAI API 429 Too Many Requests: what it means and how to fix it

What it means

Returned on inference endpoints when you send requests too frequently and reach your rate limit. Each team has per-model limits on requests per second (RPS) and tokens per minute (TPM), set by a spend-based tier, and exceeding any limit returns 429.

Causes

• Sending requests too frequently and reaching the rate limit
• Exceeding the per-model RPS or TPM cap for your team's tier
• All tokens count toward TPM, including prompt, completion, reasoning and cached prompt tokens

How to fix it

1. 1
   Fix 1.

   Implement exponential backoff to handle 429 gracefully (official OpenAI SDK example).

   Snippet

   Copy

   import os
   import time
   from openai import OpenAI, RateLimitError
   
   client = OpenAI(base_url="https://api.x.ai/v1", api_key=os.getenv("XAI_API_KEY"))
   
   def request_with_backoff(messages, max_retries=5):
       for attempt in range(max_retries):
           try:
               return client.chat.completions.create(
                   model="grok-4.7",
                   messages=messages,
               )
           except RateLimitError:
               if attempt == max_retries - 1:
                   raise
               wait = 2 ** attempt
               time.sleep(wait)

2. 2
   Fix 2.

   Reduce your request rate or increase your rate limit on xAI Console.

3. 3
   Fix 3.

   Tiers upgrade automatically based on cumulative spend; you can also submit a limit increase request through the xAI Console, or email sales@x.ai for enterprise capacity.