Grok API pricing & models

xAI · 8 live models

WandAPI serves 8 xAI models: Grok Build 0.1, Grok 4.20 Multi-Agent Beta, Grok 4.20 (Non-Reasoning), Grok 4.20, Grok 4.3, Grok 4.5, Grok 4.6, Grok 4.7. Input costs 1–2 Credits / M tokens (1 Credit = ¥1), up to 85% below list price. Compatible with OpenAI Responses; one API key for all of them, pay as you go.

Get an API KeySee pricing
Live models
8
Lowest input price
1Credits / M
Largest context
1Mtokens
Avg. 7-day uptime
—

xAI model pricing

WandAPI default channel price in Credits (1 Credit = ¥1); struck-through is the list price. Click a model for full pricing, protocols and uptime.

ModelContext · outputInput / MOutput / M7-day uptime
Grok Build 0.1grok-build-0.1256K$1.00185% off$2.002—
Grok 4.20 Multi-Agent Betagrok-4.20-multi-agent-03091M$1.251.2585% off$2.502.5—
Grok 4.20 (Non-Reasoning)grok-4.20-non-reasoning1M$1.251.2585% off$2.502.5—
Grok 4.20grok-4.20-reasoning1M$1.251.2585% off$2.502.5—
Grok 4.3grok-4.31M · 128K$1.251.2585% off$2.502.5—
Grok 4.5grok-4.5500K$2.00285% off$6.006—
Grok 4.6grok-4.6500K$2.00285% off$6.006—
Grok 4.7grok-4.7500K$2.00285% off$6.006—

Input (prompt) / Output (completion) are per 1M tokens; image and video models show their primary unit price. Uptime comes from WandAPI probes; a dash means not probed yet.

How to call xAI models

Samples use Grok Build 0.1; change the model field to call any other model from this vendor.

Projects already on the OpenAI SDK need no code changes: point base_url at https://api.wandapi.com/v1 — streaming, tool calls and usage are all mapped.

curl https://api.wandapi.com/v1/responses \
  -H "Authorization: Bearer $WANDAPI_KEY" \
  -H "content-type: application/json" \
  -d '{
    "model": "grok-build-0.1",
    "input": "Explain prompt caching in one sentence",
    "max_output_tokens": 1024
  }'

xAI model capabilities

Exactly as flagged in the catalog.

CapabilityGrok Build 0.1Grok 4.20 Multi-Agent BetaGrok 4.20 (Non-Reasoning)Grok 4.20Grok 4.3Grok 4.5Grok 4.6Grok 4.7
Vision
Function Calling
Reasoning
Prompt Caching
Response Schema
Audio
Video
File Input
Streaming
Web Search
Realtime
Session
Media
Multipart Form
Binary Response

Compare xAI models

Pricing, context, capabilities and uptime side by side.

Grok Build 0.1 vs Grok 4.20 Multi-Agent BetaGrok Build 0.1 vs Grok 4.20 (Non-Reasoning)Grok Build 0.1 vs Grok 4.20Grok Build 0.1 vs Grok 4.3Grok Build 0.1 vs Grok 4.5Grok Build 0.1 vs Grok 4.6Grok Build 0.1 vs Grok 4.7Grok 4.20 Multi-Agent Beta vs Grok 4.20 (Non-Reasoning)Grok 4.20 Multi-Agent Beta vs Grok 4.20Grok 4.20 Multi-Agent Beta vs Grok 4.3Grok 4.20 Multi-Agent Beta vs Grok 4.5Grok 4.20 Multi-Agent Beta vs Grok 4.6Grok 4.20 Multi-Agent Beta vs Grok 4.7Grok 4.20 (Non-Reasoning) vs Grok 4.20Grok 4.20 (Non-Reasoning) vs Grok 4.3Grok 4.20 (Non-Reasoning) vs Grok 4.5Grok 4.20 (Non-Reasoning) vs Grok 4.6Grok 4.20 (Non-Reasoning) vs Grok 4.7Grok 4.20 vs Grok 4.3Grok 4.20 vs Grok 4.5

FAQ

How much does the Grok API cost on WandAPI?

xAI models cost 1–2 Credits / M input tokens, with 1 Credit = ¥1; the cheapest is Grok Build 0.1. That is up to 85% below list price — see the table above.

How do I call Grok models on WandAPI with the official SDK?

Point the base URL at https://api.wandapi.com and use a WandAPI key. OpenAI Responses are supported and request bodies are identical to the official API.

Which Grok model has the largest context window?

Grok 4.20 Multi-Agent Beta, with 1M tokens. Context and max output for every model are in the pricing table above.

Can I pay for the Grok API with Alipay?

Yes. WandAPI accepts Alipay top-ups; 1 Credit = ¥1, billed per use with no monthly fee or minimum.

One API key for every xAI model

Free credits on sign-up, pay as you go, no monthly fee.

Get an API Key