gemini-4-argon

Gemini 4 Argon exposes text reasoning, multimodal understanding, structured output, function calling, and built-in tools through one generateContent endpoint. Select a capability to view its request and response shape.

Text and thinking
Multimodal understanding
Structured output
Function calling
Built-in tools

Generate content

POST /v1beta/models/gemini-4-argon:generateContent

cURL

curl --request POST \
  --url 'https://www.anyfast.ai/v1beta/models/gemini-4-argon:generateContent?key=YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "contents": [{
      "role": "user",
      "parts": [{
        "text": "Review this retry design, identify race conditions, and propose a safe implementation plan."
      }]
    }],
    "generationConfig": {
      "thinkingConfig": {"thinkingLevel": "medium"},
      "maxOutputTokens": 2048
    }
  }'

Authentication and headers

string

required

AnyFast API key.

string

default:"application/json"

required

Request media type.

Request body

object[]

required

Conversation history and the current user request. The final item must be a non-empty user turn.

string

default:"user"

Conversation role: user or model. Do not end the request with a model turn.

object[]

required

Content parts for the turn.

string

required

Text prompt or conversation content.

string

default:"medium"

Thinking level: low, medium, or high.

integer

Maximum generated tokens for a synchronous request, from 1 through 32,768.

Successful response

200

{
  "candidates": [{
    "content": {
      "parts": [{"text": "The design has a race between timeout handling and retry scheduling..."}],
      "role": "model"
    },
    "finishReason": "STOP",
    "index": 0,
    "safetyRatings": []
  }],
  "usageMetadata": {
    "promptTokenCount": 18,
    "candidatesTokenCount": 121,
    "thoughtsTokenCount": 86,
    "totalTokenCount": 225,
    "trafficType": "ON_DEMAND"
  },
  "modelVersion": "gemini-4-argon",
  "responseId": "response-example-001"
}

object[]

Generated candidates.

object

Model response turn.

string

Generated text.

string

Why generation stopped, such as STOP, MAX_TOKENS, or SAFETY.

object[]

Safety assessments returned when the candidate is evaluated.

object

Input, output, reasoning, tool, cache, and total token counts.

string

Model version that produced the response.

string

Response identifier.

Generate content

POST /v1beta/models/gemini-4-argon:generateContent

cURL

curl --request POST \
  --url 'https://www.anyfast.ai/v1beta/models/gemini-4-argon:generateContent?key=YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "contents": [{
      "role": "user",
      "parts": [
        {"text": "Summarize this architecture and identify the likely bottleneck."},
        {
          "fileData": {
            "mimeType": "image/png",
            "fileUri": "https://example.com/architecture.png"
          }
        }
      ]
    }]
  }'

Authentication and headers

string

required

AnyFast API key.

string

default:"application/json"

required

Request media type.

Request body

object[]

required

Conversation turns containing text and media.

string

default:"user"

Conversation role: user or model.

object[]

required

One or more text or media parts.

string

Instruction or question about the media.

object

Inline image, audio, video, or document data.

string

required

Media MIME type, such as image/jpeg, audio/mpeg, video/mp4, or application/pdf.

string

required

Base64-encoded media bytes.

object

Media or document supplied by URI.

string

required

Media MIME type.

string

required

Accessible media or document URI.

Successful response

200

{
  "candidates": [{
    "content": {
      "parts": [{"text": "The diagram shows a queue-backed worker tier. The database connection pool is the likely bottleneck..."}],
      "role": "model"
    },
    "finishReason": "STOP",
    "index": 0
  }],
  "usageMetadata": {
    "promptTokenCount": 846,
    "candidatesTokenCount": 94,
    "totalTokenCount": 940,
    "promptTokensDetails": [
      {"modality": "TEXT", "tokenCount": 18},
      {"modality": "IMAGE", "tokenCount": 828}
    ]
  },
  "modelVersion": "gemini-4-argon",
  "responseId": "response-example-002"
}

string

Text analysis of the supplied media.

string

Reason generation stopped.

object[]

Input token counts grouped by modality.

object[]

Output token counts grouped by modality when returned.

string

Model version that produced the response.

string

Response identifier.

Generate content

POST /v1beta/models/gemini-4-argon:generateContent

cURL

curl --request POST \
  --url 'https://www.anyfast.ai/v1beta/models/gemini-4-argon:generateContent?key=YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "contents": [{
      "role": "user",
      "parts": [{"text": "Extract the service, severity, and action: checkout-api is timing out and requires immediate rollback."}]
    }],
    "generationConfig": {
      "responseMimeType": "application/json",
      "responseSchema": {
        "type": "object",
        "properties": {
          "service": {"type": "string"},
          "severity": {"type": "string"},
          "action": {"type": "string"}
        },
        "required": ["service", "severity", "action"]
      }
    }
  }'

Authentication and headers

string

required

AnyFast API key.

string

default:"application/json"

required

Request media type.

Request body

object[]

required

Conversation turns containing the extraction or generation instruction.

string

required

Instruction and source text.

string

required

Use application/json for JSON structured output or text/x.enum for a single enum value.

object

required

Schema that constrains the generated JSON. Define explicit properties and required fields.

Successful response

200

{
  "candidates": [{
    "content": {
      "parts": [{
        "text": "{\"service\":\"checkout-api\",\"severity\":\"critical\",\"action\":\"rollback\"}"
      }],
      "role": "model"
    },
    "finishReason": "STOP",
    "index": 0
  }],
  "usageMetadata": {
    "promptTokenCount": 27,
    "candidatesTokenCount": 24,
    "totalTokenCount": 51
  },
  "modelVersion": "gemini-4-argon",
  "responseId": "response-example-003"
}

string

JSON serialized as text and conforming to the requested schema.

string

Reason generation stopped.

object

Token accounting for the request.

string

Model version that produced the response.

string

Response identifier.

Generate function call

POST /v1beta/models/gemini-4-argon:generateContent

cURL

curl --request POST \
  --url 'https://www.anyfast.ai/v1beta/models/gemini-4-argon:generateContent?key=YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "contents": [{
      "role": "user",
      "parts": [{"text": "Check the deployment status for checkout-api."}]
    }],
    "tools": [{
      "functionDeclarations": [{
        "name": "get_deployment_status",
        "description": "Return the current deployment status for a service.",
        "parameters": {
          "type": "object",
          "properties": {"service": {"type": "string"}},
          "required": ["service"]
        }
      }]
    }],
    "toolConfig": {
      "functionCallingConfig": {
        "mode": "ANY",
        "allowedFunctionNames": ["get_deployment_status"]
      }
    }
  }'

Authentication and headers

string

required

AnyFast API key.

string

default:"application/json"

required

Request media type.

Request body

object[]

required

Conversation turns containing the user request and, for later steps, previous model calls and function results.

object[]

required

Functions available to the model.

string

required

Function name.

string

When and how the model should use the function.

object

JSON Schema for function arguments.

string

default:"AUTO"

AUTO, ANY, NONE, or VALIDATED.

string[]

Subset of declared functions the model may call.

string

required

Function name when returning a tool result.

string

Call ID returned by the model.

object

Structured function result.

string

Opaque signature returned by the model. Preserve it exactly in later requests.

Function-call response

200

{
  "candidates": [{
    "content": {
      "parts": [{
        "functionCall": {
          "id": "call-example-001",
          "name": "get_deployment_status",
          "args": {"service": "checkout-api"}
        },
        "thoughtSignature": "opaque-signature-placeholder"
      }],
      "role": "model"
    },
    "finishReason": "STOP",
    "index": 0
  }],
  "usageMetadata": {
    "promptTokenCount": 86,
    "candidatesTokenCount": 18,
    "toolUsePromptTokenCount": 64,
    "totalTokenCount": 168
  },
  "modelVersion": "gemini-4-argon",
  "responseId": "response-example-004"
}

object

Function requested by the model.

string

Call ID to associate with a returned function result.

string

Function name.

object

Model-generated function arguments.

string

Opaque signature that must be preserved when the model turn is replayed.

integer

Tokens used for tool context when this metric is reported.

string

Response identifier.

Generate grounded content

POST /v1beta/models/gemini-4-argon:generateContent

cURL

curl --request POST \
  --url 'https://www.anyfast.ai/v1beta/models/gemini-4-argon:generateContent?key=YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "contents": [{
      "role": "user",
      "parts": [{"text": "Find the latest stable Gemini API model and cite the source."}]
    }],
    "tools": [{"googleSearch": {}}]
  }'

Authentication and headers

string

required

AnyFast API key.

string

default:"application/json"

required

Request media type.

Request body

object[]

required

Conversation turns containing the tool-assisted request.

object

Ground the response with current Google Search results.

object

Ground location-related answers with Google Maps data.

object

Read and analyze URLs supplied in the prompt.

object

Allow the model to execute code.

object

Allow the model to propose computer-use actions. This tool is in Preview.

string

Computer-use environment, such as ENVIRONMENT_BROWSER.

Grounded response

200

{
  "candidates": [{
    "content": {
      "parts": [{"text": "Gemini 4 Argon is the latest stable Flash model..."}],
      "role": "model"
    },
    "finishReason": "STOP",
    "groundingMetadata": {
      "webSearchQueries": ["latest stable Gemini API model"],
      "searchEntryPoint": {
        "renderedContent": "<style>...</style><div>...</div>"
      },
      "groundingChunks": [{
        "web": {
          "uri": "https://ai.google.dev/gemini-api/docs/models",
          "title": "Gemini models"
        }
      }],
      "groundingSupports": [{
        "segment": {"startIndex": 0, "endIndex": 56, "text": "Gemini 4 Argon is the latest stable Flash model"},
        "groundingChunkIndices": [0]
      }]
    }
  }],
  "usageMetadata": {
    "promptTokenCount": 15,
    "candidatesTokenCount": 58,
    "toolUsePromptTokenCount": 43,
    "totalTokenCount": 116
  },
  "modelVersion": "gemini-4-argon",
  "responseId": "response-example-005"
}

object[]

Text, executable code, code results, or other returned parts.

string[]

Search queries issued when grounding metadata is returned.

string

Google-provided HTML and CSS for the search suggestions UI when returned.

object[]

Web or Maps sources returned for the response.

object[]

Links generated text segments to grounding sources when returned.

object

Code returned when codeExecution is enabled and the model selects code execution.

object

Execution outcome and output returned after code execution.

integer

Tokens used for tool context when this metric is reported.

string

Response identifier.

Common optional request fields

object

Text-only system instruction with a parts array.

string[]

Up to five strings that stop generation when encountered.

object[]

Safety thresholds applied by harm category.

string

Supported category, such as HARM_CATEGORY_HATE_SPEECH or HARM_CATEGORY_DANGEROUS_CONTENT.

string

Blocking threshold, such as BLOCK_MEDIUM_AND_ABOVE, BLOCK_ONLY_HIGH, or BLOCK_NONE.

Error responses

integer

HTTP-compatible error code.

string

Human-readable error description.

string

Canonical error status.

A 400 response commonly indicates an invalid final conversation turn, an unsupported thinking level or tool, or an incomplete function response. A 401 response indicates an invalid API key; 429 indicates rate limiting.

Overview

Native Compatible Endpoints

Model APIs

Generate content

Authentication and headers

Request body

Successful response

Generate content

Authentication and headers

Request body

Successful response

Generate content

Authentication and headers

Request body

Successful response

Generate function call

Authentication and headers

Request body

Function-call response

Generate grounded content

Authentication and headers

Request body

Grounded response

Common optional request fields

Error responses

​Generate content

​Authentication and headers

​Request body

​Successful response

​Generate content

​Authentication and headers

​Request body

​Successful response

​Generate content

​Authentication and headers

​Request body

​Successful response

​Generate function call

​Authentication and headers

​Request body

​Function-call response

​Generate grounded content

​Authentication and headers

​Request body

​Grounded response

​Common optional request fields

​Error responses

Generate content

Authentication and headers

Request body

Successful response

Generate content

Authentication and headers

Request body

Successful response

Generate content

Authentication and headers

Request body

Successful response

Generate function call

Authentication and headers

Request body

Function-call response

Generate grounded content

Authentication and headers

Request body

Grounded response

Common optional request fields

Error responses