Agnes-2.0-Flash

Agnes-2.0-Flash is a fast and efficient language model developed by Sapiens AI, designed for agent workflows, tool calling, coding tasks, reasoning, multi-turn conversations, image understanding, and high-frequency production use cases. Agnes-2.0-Flash achieved strong performance on the Claw-Eval benchmark, ranking 9th on the General Leaderboard with a Pass^3 score of 60.9%, demonstrating strong autonomous agent capabilities among mainstream language models.

Model Overview

Agnes-2.0-Flash is optimized for fast, reliable, and cost-efficient language generation, agent task execution, and image understanding. The model supports the following capabilities:

Capability	Description
Chat Completion	Generate high-quality responses for conversations and applications
Multi-turn Conversation	Maintain context continuity across multiple turns
Image URL Input	Accept image content through publicly accessible image URLs
Image Understanding	Understand image content, analyze screenshots, and extract visual information
Tool Calling	Call external tools and functions for agent workflows
Agent Workflows	Support planning, execution, and multi-step task completion
Coding Tasks	Assist with code generation, debugging, explanation, and refactoring
Reasoning	Handle structured reasoning, task decomposition, and decision-making
Streaming Output	Return responses in real time for a better user experience
OpenAI-Compatible API	Use a request structure compatible with the OpenAI Chat Completions API

Use Cases

Agnes-2.0-Flash is suitable for the following scenarios:

Scenario	Example Use Cases
AI Assistant	General Q&A, daily assistant, productivity support
Autonomous Agents	Multi-step task execution, planning, and tool usage
Coding Assistant	Code generation, debugging, refactoring, and explanation
Workflow Automation	Task decomposition, process automation, and execution planning
Customer Support	FAQ answering, customer service chatbots, service automation
Search and Q&A	Search-based answers, summarization, information extraction
Content Generation	Marketing copy, articles, product descriptions, scripts
Developer Tools	API assistant, documentation assistant, coding copilot
AI-Native Applications	Consumer apps, productivity tools, agent applications
Image Understanding	Image description, screenshot analysis, visual Q&A, information extraction

API Information

Endpoint

Item	Description
API Endpoint	https://apihub.agnes-ai.com/v1/chat/completions
Request Method	POST
Content-Type	application/json
Authentication	Bearer Token
Authentication Header	Authorization: Bearer YOUR_API_KEY

Request Parameters

Parameter	Type	Required	Description
model	string	Yes	Model name. Use agnes-2.0-flash
messages	array	Yes	Conversation message array, including system, user, and assistant messages
messages[].content	string / array	Yes	Message content. It can be a plain text string or an array containing text and image_url content blocks
temperature	number	No	Controls output randomness. Lower values produce more deterministic results
top_p	number	No	Controls nucleus sampling. Lower values make the output more focused
max_tokens	number	No	Maximum number of tokens to generate in the response
stream	boolean	No	Whether to enable streaming output
tools	array	No	Tool definitions for tool-calling workflows
tool_choice	string / object	No	Controls whether and how the model uses tools
chat_template_kwargs	object	No	Extension field for enabling Thinking and other capabilities in OpenAI-compatible requests
thinking	object	No	Field for enabling Thinking mode in Anthropic-compatible requests

Image URL Input Support

Agnes-2.0-Flash supports image input through image URLs. Developers can pass both text instructions and an image URL in the same messages request, allowing the model to understand, analyze, answer questions about, or extract information from the image. Supported input types:

Input Type	Format	Description
Text	text	Plain text instruction or question
Image URL	image_url	Pass image content through a publicly accessible image URL

Image Content Structure

When using image URL input, messages[].content should use an array structure. Each content block represents one type of input.

{
  "role": "user",
  "content": [
    {
      "type": "text",
      "text": "Describe the content of this image."
    },
    {
      "type": "image_url",
      "image_url": {
        "url": "https://example.com/image.jpg"
      }
    }
  ]
}

Request Examples

1. Basic Chat Completion Request

Use this request to generate a standard chat completion response.

curl https://apihub.agnes-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "agnes-2.0-flash",
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful AI assistant."
      },
      {
        "role": "user",
        "content": "Explain how autonomous agents use tools to complete tasks."
      }
    ],
    "temperature": 0.7,
    "max_tokens": 1024
  }'

2. Streaming Output Request

Use this request to enable streaming output.

curl https://apihub.agnes-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "agnes-2.0-flash",
    "messages": [
      {
        "role": "user",
        "content": "Write a short product introduction for an AI assistant app."
      }
    ],
    "stream": true
  }'

3. Tool Calling Request

Use this request for agent workflows that require external tool calls.

curl https://apihub.agnes-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "agnes-2.0-flash",
    "messages": [
      {
        "role": "user",
        "content": "What is the weather like in Singapore today?"
      }
    ],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "get_weather",
          "description": "Get the current weather for a location",
          "parameters": {
            "type": "object",
            "properties": {
              "location": {
                "type": "string",
                "description": "The city and country"
              }
            },
            "required": ["location"]
          }
        }
      }
    ]
  }'

4. Image URL Input Request

Use this request to pass an image through an image URL and let the model understand or analyze the image content.

curl https://apihub.agnes-ai.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "agnes-2.0-flash",
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "Describe the content of this image."
          },
          {
            "type": "image_url",
            "image_url": {
              "url": "https://example.com/image.jpg"
            }
          }
        ]
      }
    ]
  }'

Response Format

{
  "id": "chatcmpl_xxx",
  "object": "chat.completion",
  "created": 1774432125,
  "model": "agnes-2.0-flash",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Autonomous agents use tools by understanding the user's goal, breaking it into steps, selecting the right tools, executing actions, and using the results to complete the task."
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 35,
    "completion_tokens": 58,
    "total_tokens": 93
  }
}

Response Fields

Field	Type	Description
id	string	Unique ID of the completion request
object	string	Object type, usually chat.completion
created	integer	Request timestamp
model	string	Model used for the request
choices	array	List of generated response results
choices[].index	integer	Index of the response result
choices[].message	object	Assistant message object
choices[].message.role	string	Role of the message sender
choices[].message.content	string	Response content generated by the model
choices[].finish_reason	string	Reason why generation stopped
usage	object	Token usage information
usage.prompt_tokens	integer	Number of input tokens
usage.completion_tokens	integer	Number of output tokens
usage.total_tokens	integer	Total number of tokens used

Enable Thinking for Coding Tasks

For coding, debugging, reasoning, and agent workflows, it is recommended to enable Thinking mode to improve code quality, task decomposition, and problem-solving performance.

OpenAI-Compatible Request

When using the OpenAI-compatible API format, add chat_template_kwargs.enable_thinking to the request body:

{
  "model": "agnes-2.0-flash",
  "messages": [
    {
      "role": "user",
      "content": "Help me write a Python script to process a CSV file."
    }
  ],
  "chat_template_kwargs": {
    "enable_thinking": true
  }
}

Anthropic-Compatible Request

When using the Anthropic-compatible API format, add the thinking field to the request body:

{
  "model": "agnes-2.0-flash",
  "messages": [
    {
      "role": "user",
      "content": "Help me refactor this TypeScript function and explain the changes."
    }
  ],
  "thinking": {
    "type": "enabled",
    "budget_tokens": 2048
  }
}

budget_tokens controls the maximum Thinking token budget. For common coding tasks, it is recommended to start with 2048. For more complex debugging, refactoring, or multi-step agent tasks, you can increase the value as needed.

Features and Compatibility

Agnes-2.0-Flash supports the following capabilities:

Chat Completion
Multi-turn conversation
System prompt
Image URL input
Image understanding
Streaming output
Tool calling
Agent workflows
Coding tasks
Reasoning tasks
JSON-style output
OpenAI Chat Completions API-compatible request structure

Best Practices

Prompt Writing Tips

For better results, provide clear instructions, sufficient context, and the expected output format.

Example: Product Copy Generation

text

You are a product marketing expert. Write a concise App Store description for an AI assistant app. The tone should be clear, professional, and user-friendly.

Example: Coding Task

For coding tasks, provide the programming language, framework, error message, and expected behavior.

text

Help me debug this React component. The issue is that the button state does not update after clicking. Explain the cause and provide the corrected code.

Example: Agent Workflow

For agent workflows, clearly describe the goal, available tools, and task constraints.

text

You are an autonomous research agent. Search for relevant information, summarize the key findings, and return the result in a structured format with source links.

Example: Image Understanding Task

For image understanding tasks, clearly state what the model should focus on, such as overall description, text extraction, UI analysis, object recognition, or structured output.

text

Analyze this screenshot. Identify the main UI elements, explain the possible issue, and provide suggestions to improve the user experience.

Recommended Prompt Structure

Use the following structure to organize prompts:

text

[Role] + [Task] + [Context] + [Requirements] + [Output Format]

Example

text

You are a senior product manager. Analyze this feature idea for an AI assistant app. Consider user value, implementation complexity, risks, and return the result in a structured table.

Image Understanding Prompt Example

text

You are an image analysis assistant. Analyze the provided image URL, summarize the key information, identify potential issues, and return the result in a structured table.

Image URL Usage Tips

The image URL must be publicly accessible.
If the image URL requires login, authentication, or has hotlink protection, the model may not be able to read it.
It is recommended to use standard image formats such as JPG, JPEG, PNG, or WebP.
For screenshots, error images, or product UI images, add text instructions to clarify what the model should focus on.
Image URL input can be used together with tool calling, streaming output, and agent workflows.

Model Limits

Item	Value
Context	256K
Max Output	65.5K

Pricing

Type	Price	Current Price
Input Tokens	$0.03 / 1M tokens	$0 / 1M tokens
Output Tokens	$0.15 / 1M tokens	$0 / 1M tokens

Notes

Use agnes-2.0-flash as the model name.
A basic Chat Completion request must include model and messages.
messages[].content can be a plain text string or an array containing text and an image URL.
To input an image, use image_url and provide a publicly accessible image URL.
To enable streaming responses, set stream to true.
For tool-calling workflows, provide tools and optionally tool_choice.
temperature controls randomness. Lower values are better for deterministic tasks, while higher values are better for creative generation.
Agnes-2.0-Flash is suitable for production applications that require fast responses, strong task completion, image understanding, and reliable agent performance.

​Agnes-2.0-Flash

​Model Overview

​Use Cases

​API Information

​Endpoint

​Request Parameters

​Image URL Input Support

​Image Content Structure

​Request Examples

​1. Basic Chat Completion Request

​2. Streaming Output Request

​3. Tool Calling Request

​4. Image URL Input Request

​Response Format

​Response Fields

​Enable Thinking for Coding Tasks

​OpenAI-Compatible Request

​Anthropic-Compatible Request

​Features and Compatibility

​Best Practices

​Prompt Writing Tips

​Example: Product Copy Generation

​Example: Coding Task

​Example: Agent Workflow

​Example: Image Understanding Task

​Recommended Prompt Structure

​Example

​Image Understanding Prompt Example

​Image URL Usage Tips

​Model Limits

​Pricing

​Notes

Agnes-2.0-Flash

Model Overview

Use Cases

API Information

Endpoint

Request Parameters

Image URL Input Support

Image Content Structure

Request Examples

1. Basic Chat Completion Request

2. Streaming Output Request

3. Tool Calling Request

4. Image URL Input Request

Response Format

Response Fields

Enable Thinking for Coding Tasks

OpenAI-Compatible Request

Anthropic-Compatible Request

Features and Compatibility

Best Practices

Prompt Writing Tips

Example: Product Copy Generation

Example: Coding Task

Example: Agent Workflow

Example: Image Understanding Task

Recommended Prompt Structure

Example

Image Understanding Prompt Example

Image URL Usage Tips

Model Limits

Pricing

Notes