Grok 3

Developer documentation

Grok 3 for text generation, reasoning, tool calling, and streaming responses.

Model Reference

Chat and reasoning models

Language, code, reasoning, multimodal chat, tool calling, and streaming responses. Endpoint: http://omixa.cloud/api/v1/chat/completions

Grok 3

grok-3

Grok 3 for text generation, reasoning, tool calling, and streaming responses.

Chat Streaming Tools Context window: 131,072 tokens Max output: 8,192 tokens

input per 1m tokens $3.000000

output per 1m tokens $15.000000

minimum hold $0.010000

Integration reference

Connect Grok 3

Use Omixa's unified endpoint and your workspace API key. Provider routing, billing, failover, and usage records are handled by Omixa.

POST http://omixa.cloud/api/v1/chat/completions

Provider: azure-foundry
Endpoint type: azure_model_inference_chat
Context window: 131,072 tokens
Max output: 8,192 tokens
Streaming supported
Tool/function calling supported

Request schema

Request fields

Only send options supported by this model. Required fields and accepted values are listed below.

Field	Type	Required	Accepted values	Description
`model`	string	Yes	grok-3	Use `grok-3`. Omixa resolves the active provider route and failover key automatically.
`messages`	array	Yes	Any valid value	Conversation turns. Each item contains `role` and `content`. Roles are `system`, `user`, `assistant`, and tool roles when supported.
`messages[].content`	string\|array	Yes	Any valid value	Text prompt or multimodal parts. Multimodal-capable models accept `{type:"text"}`, `{type:"image_url"}`, and inline files supported by Omixa.
`stream`	boolean	No	Any valid value	When true, returns OpenAI-compatible Server-Sent Events until `data: [DONE]`.
`max_tokens\|max_completion_tokens`	integer	No	Any valid value	Maximum answer tokens. Omixa maps the correct field for Google, Azure chat, and reasoning model families.
`temperature`	number	No	Any valid value	Sampling randomness for non-reasoning models. Reasoning-only models may ignore or reject it, so Omixa strips it where needed.
`top_p`	number	No	Any valid value	Nucleus sampling. Use either `temperature` or `top_p` for predictable tuning.
`stop`	string\|array	No	Any valid value	Stop sequence or sequences. Not all provider routes support it.
`response_format`	object	No	Any valid value	Structured output hints such as `{ "type": "json_object" }` when the selected provider supports JSON mode.
`tools`	array	No	Any valid value	OpenAI-style tool/function declarations. Omixa forwards compatible declarations to tool-capable providers.
`tool_choice`	string\|object	No	Any valid value	`auto`, `none`, `required`, or a specific tool choice when the provider supports it.

Ready to send

Payload and response

Start with this model-safe payload and expect the normalized Omixa response shape shown beside it.

Example JSON payload

{
    "model": "grok-3",
    "messages": [
        {
            "role": "system",
            "content": "You are a concise API assistant."
        },
        {
            "role": "user",
            "content": "Explain this model in three practical bullet points."
        }
    ],
    "stream": true,
    "max_tokens": 2048,
    "temperature": 0.7,
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "lookup_order",
                "description": "Return order status by ID.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "order_id": {
                            "type": "string"
                        }
                    },
                    "required": [
                        "order_id"
                    ]
                }
            }
        }
    ],
    "tool_choice": "auto"
}

Response shape

{
    "object": "chat.completion",
    "choices": [
        {
            "message": {
                "role": "assistant",
                "content": "..."
            }
        }
    ],
    "usage": {
        "prompt_tokens": 123,
        "completion_tokens": 45,
        "total_tokens": 168
    }
}

Language examples

Copy-ready integration code

Replace the example API key with a workspace key and keep model-specific fields unchanged unless the table above marks them optional.

cURL

curl -X POST http://omixa.cloud/api/v1/chat/completions \
  -H "Authorization: Bearer omx_live_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-3",
    "messages": [
        {
            "role": "system",
            "content": "You are a concise API assistant."
        },
        {
            "role": "user",
            "content": "Explain this model in three practical bullet points."
        }
    ],
    "stream": true,
    "max_tokens": 2048,
    "temperature": 0.7,
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "lookup_order",
                "description": "Return order status by ID.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "order_id": {
                            "type": "string"
                        }
                    },
                    "required": [
                        "order_id"
                    ]
                }
            }
        }
    ],
    "tool_choice": "auto"
}'

JavaScript fetch

const response = await fetch('http://omixa.cloud/api/v1/chat/completions', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer omx_live_xxx',
    'Content-Type': 'application/json'
  },
  body: "{\n    \"model\": \"grok-3\",\n    \"messages\": [\n        {\n            \"role\": \"system\",\n            \"content\": \"You are a concise API assistant.\"\n        },\n        {\n            \"role\": \"user\",\n            \"content\": \"Explain this model in three practical bullet points.\"\n        }\n    ],\n    \"stream\": true,\n    \"max_tokens\": 2048,\n    \"temperature\": 0.7,\n    \"tools\": [\n        {\n            \"type\": \"function\",\n            \"function\": {\n                \"name\": \"lookup_order\",\n                \"description\": \"Return order status by ID.\",\n                \"parameters\": {\n                    \"type\": \"object\",\n                    \"properties\": {\n                        \"order_id\": {\n                            \"type\": \"string\"\n                        }\n                    },\n                    \"required\": [\n                        \"order_id\"\n                    ]\n                }\n            }\n        }\n    ],\n    \"tool_choice\": \"auto\"\n}"
});
const data = await response.json();

Python requests

import requests

response = requests.post(
    'http://omixa.cloud/api/v1/chat/completions',
    headers={'Authorization': 'Bearer omx_live_xxx', 'Content-Type': 'application/json'},
    json={
    "model": "grok-3",
    "messages": [
        {
            "role": "system",
            "content": "You are a concise API assistant."
        },
        {
            "role": "user",
            "content": "Explain this model in three practical bullet points."
        }
    ],
    "stream": True,
    "max_tokens": 2048,
    "temperature": 0.7,
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "lookup_order",
                "description": "Return order status by ID.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "order_id": {
                            "type": "string"
                        }
                    },
                    "required": [
                        "order_id"
                    ]
                }
            }
        }
    ],
    "tool_choice": "auto"
}
)
print(response.json())

PHP cURL

$ch = curl_init('http://omixa.cloud/api/v1/chat/completions');
curl_setopt_array($ch, [
    CURLOPT_POST => true,
    CURLOPT_HTTPHEADER => ['Authorization: Bearer omx_live_xxx', 'Content-Type: application/json'],
    CURLOPT_POSTFIELDS => '{
    "model": "grok-3",
    "messages": [
        {
            "role": "system",
            "content": "You are a concise API assistant."
        },
        {
            "role": "user",
            "content": "Explain this model in three practical bullet points."
        }
    ],
    "stream": true,
    "max_tokens": 2048,
    "temperature": 0.7,
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "lookup_order",
                "description": "Return order status by ID.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "order_id": {
                            "type": "string"
                        }
                    },
                    "required": [
                        "order_id"
                    ]
                }
            }
        }
    ],
    "tool_choice": "auto"
}',
    CURLOPT_RETURNTRANSFER => true,
]);
$response = curl_exec($ch);

C# HttpClient

using var client = new HttpClient();
client.DefaultRequestHeaders.Authorization = new System.Net.Http.Headers.AuthenticationHeaderValue("Bearer", "omx_live_xxx");
var json = @"{
    ""model"": ""grok-3"",
    ""messages"": [
        {
            ""role"": ""system"",
            ""content"": ""You are a concise API assistant.""
        },
        {
            ""role"": ""user"",
            ""content"": ""Explain this model in three practical bullet points.""
        }
    ],
    ""stream"": true,
    ""max_tokens"": 2048,
    ""temperature"": 0.7,
    ""tools"": [
        {
            ""type"": ""function"",
            ""function"": {
                ""name"": ""lookup_order"",
                ""description"": ""Return order status by ID."",
                ""parameters"": {
                    ""type"": ""object"",
                    ""properties"": {
                        ""order_id"": {
                            ""type"": ""string""
                        }
                    },
                    ""required"": [
                        ""order_id""
                    ]
                }
            }
        }
    ],
    ""tool_choice"": ""auto""
}";
var response = await client.PostAsync("http://omixa.cloud/api/v1/chat/completions", new StringContent(json, System.Text.Encoding.UTF8, "application/json"));
var body = await response.Content.ReadAsStringAsync();

Go net/http

payload := []byte(`{
    "model": "grok-3",
    "messages": [
        {
            "role": "system",
            "content": "You are a concise API assistant."
        },
        {
            "role": "user",
            "content": "Explain this model in three practical bullet points."
        }
    ],
    "stream": true,
    "max_tokens": 2048,
    "temperature": 0.7,
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "lookup_order",
                "description": "Return order status by ID.",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "order_id": {
                            "type": "string"
                        }
                    },
                    "required": [
                        "order_id"
                    ]
                }
            }
        }
    ],
    "tool_choice": "auto"
}`)
req, _ := http.NewRequest("POST", "http://omixa.cloud/api/v1/chat/completions", bytes.NewReader(payload))
req.Header.Set("Authorization", "Bearer omx_live_xxx")
req.Header.Set("Content-Type", "application/json")
resp, err := http.DefaultClient.Do(req)

Streaming note

Set `"stream": true` and read Server-Sent Events line by line until `data: [DONE]`. Each event is OpenAI-compatible `chat.completion.chunk` JSON.

Production checklist

Operational notes

Authenticate with `Authorization: Bearer omx_live_xxx`.
Omixa handles provider keys, routing, billing, failover, and usage recording behind this endpoint.