Skip to main content
The Info API provides endpoints to discover available models and modes, including their capabilities, pricing, and configuration details.

​
List Models

Retrieve information about all available AI models, including their capabilities, pricing, and limits. 
GET https://api.sup.ai/v1/info/models

​
Response

Returns a JSON object with text and image model categories: 
{
  "text": {
    "anthropic/claude-sonnet-4.5": {
      "name": "Claude Sonnet 4.5",
      "shortName": "Sonnet 4.5",
      "description": "Our go-to model for sophisticated work...",
      "contextWindow": 200000,
      "maxOutputTokens": 64000,
      "fileInputLimit": 20,
      "features": ["tools", "pdf-input", "thinking", "image-input"],
      "deprecated": null,
      "nanoDollarsPerMillionTokens": {
        "input": "3000000000",
        "inputCacheRead": "300000000",
        "inputCacheWrite": "3750000000",
        "output": "15000000000",
        "reasoning": null
      },
      "nanoDollarsPerMillionToolCalls": { "webSearch": "10000000000000" }
    }
  },
  "image": {
    "openai/gpt-image-1": {
      "name": "GPT Image 1",
      "description": "Image generation model...",
      "features": ["image-output"]
    }
  }
}

​
Model Properties

​
name
string
Full display name of the model.
​
shortName
string | null
Abbreviated display name, if available.
​
description
string
User-facing description of the model’s capabilities and best use cases.
​
contextWindow
number
Maximum context window size in tokens. Example: 200000 for 200K tokens.
​
maxOutputTokens
number
Maximum number of tokens the model can generate in a single response.
​
fileInputLimit
number
Maximum number of files that can be included in a single request.
​
features
array
List of features supported by this model:
FeatureDescription
toolsFunction/tool calling support
thinkingExtended reasoning capabilities
image-inputCan process images in messages
pdf-inputCan process PDFs
image-outputCan generate images
​
deprecated
object | null
Deprecation information if the model is deprecated:
{
  "at": "2024-01-15T00:00:00Z",
  "reason": "Replaced by newer version",
  "replacement": "anthropic/claude-sonnet-4.5"
}
​
nanoDollarsPerMillionTokens
object
Pricing in nano-dollars (1 billionth of a dollar) per million tokens:
FieldDescription
inputCost per million input tokens
inputCacheReadCost per million cached input tokens (if supported)
inputCacheWriteCost per million tokens written to cache (if supported)
outputCost per million output tokens
reasoningCost per million reasoning tokens (if separate billing)
​
nanoDollarsPerMillionToolCalls
object
Pricing for tool calls:
FieldDescription
webSearchCost per million web search calls (if supported)

​
Example

curl https://api.sup.ai/v1/info/models \
  -H "Authorization: Bearer YOUR_API_KEY"

​
List Modes

Retrieve information about all available chat modes, including their model counts, thinking effort, and descriptions. 
GET https://api.sup.ai/v1/info/modes

​
Response

Returns a JSON object mapping mode IDs to their configuration: 
{
  "auto": {
    "name": "Auto",
    "description": "Intelligently selects the optimal mode for your task...",
    "models": 1,
    "rank": -1,
    "thinkingEffort": "minimal",
    "thinkingBudget": 256
  },
  "fast": {
    "name": "Fast",
    "description": "Instant responses for trivial tasks...",
    "models": 1,
    "rank": 0,
    "thinkingEffort": "minimal",
    "thinkingBudget": 256
  },
  "thinking": {
    "name": "Thinking",
    "description": "The default mode for most development work...",
    "models": 3,
    "rank": 1,
    "thinkingEffort": "medium",
    "thinkingBudget": 2048
  },
  "deep-thinking": {
    "name": "Deep Thinking",
    "description": "Advanced problem-solving for complex challenges...",
    "models": 6,
    "rank": 2,
    "thinkingEffort": "high",
    "thinkingBudget": 3072
  },
  "pro": {
    "name": "Pro",
    "description": "Maximum rigor for high-stakes, mission-critical work...",
    "models": 9,
    "rank": 3,
    "thinkingEffort": "extra-high",
    "thinkingBudget": 4096
  },
  "image": {
    "name": "Image",
    "description": "Visual content creation and editing...",
    "models": 1,
    "rank": 4,
    "thinkingEffort": "minimal",
    "thinkingBudget": 256
  }
}

​
Mode Properties

​
name
string
Display name of the mode.
​
description
string
User-facing description of when to use this mode.
​
models
number
Number of AI models used in parallel for this mode.
​
rank
number
Priority rank of the mode. Higher values indicate more powerful modes. The auto mode has rank -1 as it dynamically selects other modes.
​
thinkingEffort
string
The thinking effort level for this mode:
LevelDescription
minimalQuick, direct responses with minimal deliberation
lowLight reasoning for straightforward tasks
mediumBalanced thinking for standard work
highDeep reasoning for complex problems
extra-highMaximum reasoning for mission-critical tasks
​
thinkingBudget
number
Token budget allocated for extended thinking. Models with thinking capabilities use this budget to reason through problems before responding.
EffortBudget
minimal256 tokens
low1,024 tokens
medium2,048 tokens
high3,072 tokens
extra-high4,096 tokens

​
Example

curl https://api.sup.ai/v1/info/modes \
  -H 'Authorization: Bearer YOUR_API_KEY'

​
Mode Selection Guide

Best for: Most use casesThe auto mode intelligently analyzes your prompt and selects the optimal mode. It starts with minimal resources and escalates when complexity is detected. This is the recommended default for most applications.
Best for: Trivial tasks, typo fixes, simple formattingSingle model with minimal thinking. Use when speed is critical and the task is straightforward. Examples: spell checking, simple code formatting, basic calculations.
Best for: Standard development workThree models with medium thinking effort. The default for most coding tasks, debugging, feature implementation, and analysis. Balances quality with response time.
Best for: Complex multi-faceted problemsSix models with high thinking effort. Use for architectural decisions, system design, complex debugging, and problems requiring parallel exploration of multiple approaches.
Best for: High-stakes, mission-critical workNine models with extra-high thinking effort. Reserved for legal, financial, medical, or regulatory work where maximum accuracy and verification are essential.
Best for: Visual content creationSpecialized mode for generating and editing images. Optimized for interpreting visual descriptions and producing high-quality imagery.

​
Error Responses

StatusDescription
401Unauthorized - missing or invalid API key