Model specifications
- Context
- 500,000 tokens
- Maximum output
- 450,000 tokens
- Architecture
- text+image+file->text
- Tokenizer
- Grok
- Knowledge cutoff
- Not provided
- Moderated
- No
Complete pricing
Synchronized OpenRouter rates. Token prices are shown per one million tokens.
- Input
- $1.6 per 1M tokens
- ≤200K$1.6
- >200K$3.2
- Output
- $4.8 per 1M tokens
- ≤200K$4.8
- >200K$9.6
- Cache read
- $0.4 per 1M tokens
- ≤200K$0.4
- >200K$0.8
- Web search
- $5 per 1,000 calls
Model configuration
Default parameters
temperature- 0.7
top_p- 0.95
Reasoning
- Reasoning required
- Yes
- Default parameters
- high
- Capabilities
- xhigh, high, medium, low
Quick start
Set OPENROUTER_API_KEY locally. Python requires requests; JavaScript runs in Node.js. Keep the key on the server.
This submits an asynchronous batch using the base model ID. A job ID confirms submission, not completion.
curl --fail-with-body https://openrouter.ai/api/beta/batches \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"endpoint":"/v1/chat/completions","model":"x-ai/grok-4.7","requests":[{"custom_id":"request-1","body":{"messages":[{"role":"user","content":"Hello!"}]}}]}'
# Replace JOB_ID with the returned id to check progress.
curl --fail-with-body "https://openrouter.ai/api/beta/batches/JOB_ID" \
-H "Authorization: Bearer $OPENROUTER_API_KEY"Input → Output
Yes · xhigh · high · medium · low
Supported API parameters
include_reasoninglogprobsmax_tokensreasoningreasoning_effortresponse_formatseedstructured_outputstemperaturetool_choicetoolstop_logprobstop_p1 Provider
Checked: November 4, 2026
Provider availability, latency, throughput and routing can change continuously. Open the source page for current operational data.
xAI
Not provided- Context
- 500K
- Maximum output
- 450K
- Input
- $1.6
- Output
- $4.8
- Cache read
- $0.4
- Cache write
- Not provided