Model introduction
Grok 4.5 is xAI's flagship model for code and general tasks, with agentic tool calling and configurable reasoning.
Model capabilities
Specifications come from the cited model material. Catalog tags help identify access features; use the source documentation for exact limits.
Verified specifications
- Official positioning
- Flagship model for code and general tasks
- Context window
- 500,000 tokens
- Verified capabilities
- Agentic tool calling; configurable reasoning
Sources: xAI Grok 4.5 model documentation
Independent capability evaluation
Only public measurements matched to the exact model ID and reasoning profile are shown; nearby variants are not substituted.
Intelligence index
54
Cohort rank #9 / 186
Output speed
69.2 tok/s
Cohort rank #87 / 186
First-token latency
10.71s
Reasoning, high effort
A strong composite result with middle-range speed and latency for coding and general tasks with tunable reasoning.
Evaluation profile: Reasoning, high effort. Artificial Analysis · Methodology
Model pricing and access
Prices are read directly from the MoleAPI console catalog and shown by current billing group and context tier.
Your final charge follows the account group shown in the console.
Live pricing source
Pricing table
USD / 1M tokensStandardx1defaultInput $2 · Output $6
Input
$2 / 1M tokens
Output
$6 / 1M tokens
Cache read
$0.5 / 1M tokens
Discountx0.8discountInput $1.6 · Output $4.8
Input
$1.6 / 1M tokens
Output
$4.8 / 1M tokens
Cache read
$0.4 / 1M tokens
Relayx0.3relayInput $0.6 · Output $1.8
Input
$0.6 / 1M tokens
Output
$1.8 / 1M tokens
Cache read
$0.15 / 1M tokens
Temporary (availability may vary)x0.1tempInput $0.2 · Output $0.6
Input
$0.2 / 1M tokens
Output
$0.6 / 1M tokens
Cache read
$0.05 / 1M tokens
Access protocols
Available billing groups: Standard, Discount, Relay, Temporary (availability may vary)
- openai
- POST
- /v1/chat/completions
- openai-response
- POST
- /v1/responses
- openai-response-compact
- POST
- /v1/responses/compact
- anthropic
- POST
- /v1/messages
- gemini
- POST
- /v1beta/models/{model}:generateContent
- openai-alpha-search
- POST
- /v1/alpha/search
Complete model guide
Prepared from vendor sources and independently reviewed. · 2026-07-28
Grok 4.5 model guide
Grok 4.5 is xAI’s frontier model for coding, agentic tasks, and knowledge work, with training data spanning science, engineering, and mathematics. It supports text and image inputs, reasoning, function calling, and structured outputs. [1] [9]
Positioning and core capabilities
Grok 4.5 is positioned for agentic software development, engineering, scientific, mathematical, and workflow-oriented tasks. [1] [9] Its configurable reasoning effort supports low, medium, and high settings, with high set as the default. [9]
The model supports function calling alongside web search, X search, and code-execution tools in xAI’s documented API environment. [9] Structured-output modes include JSON Schema, JSON object, and text responses. [7]
Capabilities and specs
Grok 4.5 combines long-context text processing with image input and tool-oriented output capabilities. [1] Its knowledge cutoff is February 1, 2026. [9]
| Specification | Details |
|---|---|
| Input modalities | Text and image [1] |
| Output modality | Text [1] |
| Context window | 500,000 tokens [1] |
| Reasoning | Configurable low, medium, or high effort [9] |
| Tool integration | Function calling, web search, X search, and code execution [9] |
| Structured output | JSON Schema, JSON object, and text modes [7] |
For supported JSON Schema features, xAI documents schema-conformant structured responses and tool-call arguments. [7]
Use cases
Grok 4.5 is designed for coding and agentic development workflows that combine reasoning with function calls and external tools. [1] [9] Its science, engineering, and mathematics positioning also suits technical analysis and knowledge-work tasks that benefit from long-context document handling or image input. [1] [9]
Structured outputs support applications that require machine-readable responses for extraction, classification, workflow routing, and tool invocation. [7]
Material limitations
Grok 4.5 has a 500,000-token context window, which defines the maximum documented context capacity for an individual request. [1] Its knowledge cutoff is February 1, 2026. [9]
Structured-output guarantees apply to supported JSON Schema features; unsupported or over-limit schema constraints are handled on a best-effort basis. [7]
Sources
The following xAI documentation supports the model positioning, specifications, and capability details in this guide.
Code examples
These examples use MoleAPI's Responses API endpoint and run after you replace the API key.
curl https://api.moleapi.com/v1/responses \
-H "Authorization: Bearer $MOLEAPI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "grok-4.5",
"input": "Explain this problem step by step."
}'Frequently asked questions
How is grok-4.5 priced?
This page reads prices from the MoleAPI console API and updates with model prices, context tiers, and account groups.
How can I access grok-4.5?
Protocols and endpoints come from the supported_endpoint_types field in the MoleAPI model catalog.
How do I switch an existing project to grok-4.5?
Keep the MoleAPI API address and key, replace the model parameter with the model ID on this page, then check protocol-specific parameter differences.
Where do the grok-4.5 model details come from?
Capabilities and limitations are checked against xAI and the other cited pages. Pricing and available protocols come only from the MoleAPI console.