Grok 4.5 Multimodal Reasoning: AI Large Language Models
Analyze images, solve logic puzzles, and browse the live web with Grok 4.5 on Muapi. High-performance reasoning. Try free, pay per generation.
Analyze images, solve logic puzzles, and browse the live web with Grok 4.5 on Muapi. High-performance reasoning. Try free, pay per generation.
About this model
Grok 4.5 is xAI’s powerful conversational reasoning model, engineered to deliver advanced multimodal analysis, adjustable reasoning depth, and built-in web search capabilities. It accepts mixed text and image inputs and supports structured JSON outputs, making it ideal for software development, technical research, and real-time knowledge queries. On Muapi, the standard async (/grok-4-5) and live streaming (/grok-4-5/stream) endpoints are billed per token, saving you up to 50% compared to typical API rates.
Cost analysis
| Provider | Cost | Notes |
|---|---|---|
| muapiapp | $1.60/M input tokens, $4.80/M output tokens | Token-based billing. Minimum $0.0001 per call. Up to 50% cheaper than direct enterprise subscriptions. |
| xAI (official) | $1.60/M input tokens, $4.80/M output tokens | Standard retail pricing tier. |
Token-based billing. Minimum $0.0001 per call. Up to 50% cheaper than direct enterprise subscriptions.
Standard retail pricing tier.
** Competitor pricing is estimated based on similar model architectures and usage tiers.
Configuration schema
| Parameter | Type | Description | Default |
|---|---|---|---|
| Prompt | string | The user message or instruction. | Solve this complex math puzzle step by step. |
| Image URL | string | Optional image URL for multimodal requests. | undefined |
| System Prompt | string | Optional system-level instruction to guide model behavior. | You are an expert tutor. |
| Reasoning Effort | Enum (4 options) | Reasoning effort level. Higher values provide more thorough reasoning. | low |
| Web Search | boolean | Enable web search tool for real-time information retrieval. | false |
The user message or instruction.
Solve this complex math puzzle step by step.Optional image URL for multimodal requests.
undefinedOptional system-level instruction to guide model behavior.
You are an expert tutor.Reasoning effort level. Higher values provide more thorough reasoning.
lowEnable web search tool for real-time information retrieval.
falseDeveloper documentation
Submit prompts using standard async or real-time streaming Server-Sent Events (SSE).
POST /api/v1/grok-4-5
Returns a request_id immediately. Poll /api/v1/predictions/{id}/result for completion status.
curl -X POST https://muapi.ai/api/grok-4-5 \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "Solve this complex logic puzzle step by step.",
"reasoning_effort": "high"
}'
POST /api/v1/grok-4-5/stream
Returns a chunked Content-Type: text/event-stream stream of server-sent events, ending with data: [DONE].
Frequently asked
Yes, Grok 4.5 is a multimodal model. You can supply an optional image_url in the request payload to analyze screenshots, graphs, or photographs.
Billing is token-based: input tokens cost $1.60 per million, and output tokens cost $4.80 per million. The minimum cost per request is $0.0001. Charges are automatically deducted from your account wallet balance.
Yes. If web_search is set to true in the payload, Grok 4.5 utilizes built-in search tools to fetch the latest real-time information from the web.
The standard route runs asynchronously and returns a request ID, making it ideal for background processing and scripts. The streaming route sends Server-Sent Events (SSE) in real-time, making it best for conversational chat interfaces.