invokeable()
← directory / providers / xai

xAI API

first-party
10
models served
XAI_API_KEY
auth
api.x.ai/v1
base url

OpenAI-compatible (also Anthropic-compatible endpoint available). docs →

01Request template

Every model below is invoked with this shape — swap in the provider model ID from the table.

shell · template
curl https://api.x.ai/v1/chat/completions \
  -H "Authorization: Bearer $XAI_API_KEY" \
  -H "content-type: application/json" \
  -d '{
    "model": "{{MODEL_ID}}",
    "messages": [{"role": "user", "content": "Hello, world"}]
  }'

02Models served · newest first

ModelProvider model IDIn $/MOut $/MContext
Grok 4.6xAI grok-4.6 500K
Grok 4.5xAI grok-4.5 500K
Grok 4.3xAI grok-4.3 1.25 2.50 1M
Grok Build 0.1xAI grok-build-0.1 256K
Grok 4.20 (Non-Reasoning)xAI grok-4.20-0309-non-reasoning 1M
Grok 4.20 (Reasoning)xAI grok-4.20-0309-reasoning 1M
Grok 4.1 FastxAI grok-4.1-fast 0.20 0.50 2M
Grok 4 FastxAI grok-4-fast 0.20 0.50 2M
Grok Code Fast 1xAI grok-code-fast-1 0.20 1.50 256K
Grok 4xAI grok-4 3.00 15.00 256K

Dimmed prices are canonical — this provider's own quote isn't captured yet.