Model Management

Query available AI models and their capabilities across all supported providers.

List Models

Returns the list of available AI models across all providers (Claude, Groq, vLLM).

Endpoint: https://apis.threatwinds.com/api/ai/v1/models

Method: GET

Parameters

Headers

Header	Type	Required	Description
Authorization	string	Optional*	Bearer token for session authentication
api-key	string	Optional*	API key for key-based authentication
api-secret	string	Optional*	API secret for key-based authentication

Note: You must use either Authorization header OR API key/secret combination.

Request

To list all available models, use a GET request:

curl -X 'GET' \
  'https://apis.threatwinds.com/api/ai/v1/models' \
  -H 'accept: application/json' \
  -H 'Authorization: Bearer <token>'

Or using API key and secret:

curl -X 'GET' \
  'https://apis.threatwinds.com/api/ai/v1/models' \
  -H 'accept: application/json' \
  -H 'api-key: your-api-key' \
  -H 'api-secret: your-api-secret'

Response

A successful response will return all available models across all providers.

Success Response (200 OK)

{
  "object": "list",
  "data": [
    {
      "id": "claude-sonnet-4",
      "object": "model",
      "name": "Claude Sonnet 4.5",
      "provider": "claude",
      "owned_by": "Anthropic PBC",
      "created": 2025,
      "capabilities": [
        "chat",
        "tools-use",
        "reasoning",
        "code-generation",
        "image"
      ],
      "limits": {
        "max_input_tokens": 64000,
        "max_completion_tokens": 64000,
        "max_total_tokens": 128000
      }
    },
    {
      "id": "claude-opus-4",
      "object": "model",
      "name": "Claude Opus 4.5",
      "provider": "claude",
      "owned_by": "Anthropic PBC",
      "created": 2025,
      "capabilities": [
        "chat",
        "tools-use",
        "reasoning",
        "code-generation",
        "image"
      ],
      "limits": {
        "max_input_tokens": 32000,
        "max_completion_tokens": 32000,
        "max_total_tokens": 64000
      }
    },
    {
      "id": "gpt-oss-20b",
      "object": "model",
      "name": "GPT OSS 20B",
      "provider": "groq",
      "owned_by": "OpenAI",
      "created": 2025,
      "capabilities": [
        "chat",
        "code-generation",
        "tools-use",
        "reasoning"
      ],
      "limits": {
        "max_input_tokens": 131072,
        "max_completion_tokens": 65536,
        "max_total_tokens": 196608
      }
    }
  ]
}

Response Schema

Field	Type	Description
object	string	Response type, always “list”
data	array	List of model details
data[].id	string	Model unique identifier
data[].object	string	Object type, always “model”
data[].name	string	Human-readable model name
data[].provider	string	Provider identifier: claude, groq, vllm
data[].owned_by	string	Organization that owns the model
data[].created	integer	Year of model release
data[].capabilities	array	List of model capabilities
data[].limits	object	Token limit information
limits.max_input_tokens	integer	Maximum input tokens
limits.max_completion_tokens	integer	Maximum completion tokens
limits.max_total_tokens	integer	Maximum total tokens (input + completion)

Model Capabilities

Capability	Description
chat	Text-based conversation
text-generation	General text generation
code-generation	Code generation and completion
tools-use	Function/tool calling support
reasoning	Extended reasoning capabilities
image	Image understanding (vision)
audio	Audio processing
video	Video processing

Business Logic

• Retrieves models from in-memory provider registry
• Models aggregated across all registered providers
• No database queries - data from provider configuration
• Model list may change as providers update their offerings

Error Codes

Status Code	Description	Possible Cause
200	OK	Request successful
400	Bad Request	Invalid request
401	Unauthorized	Missing or invalid authentication
403	Forbidden	Insufficient permissions

---

Get Model Details

Returns detailed information about a specific model.

Endpoint: https://apis.threatwinds.com/api/ai/v1/models/{id}

Method: GET

Parameters

Headers

Header	Type	Required	Description
Authorization	string	Optional*	Bearer token for session authentication
api-key	string	Optional*	API key for key-based authentication
api-secret	string	Optional*	API secret for key-based authentication

Note: You must use either Authorization header OR API key/secret combination.

Path Parameters

Parameter	Type	Required	Description	Example
id	string	Yes	Model identifier	claude-sonnet-4

Request

To get details for a specific model, use a GET request:

curl -X 'GET' \
  'https://apis.threatwinds.com/api/ai/v1/models/claude-sonnet-4' \
  -H 'accept: application/json' \
  -H 'Authorization: Bearer <token>'

Or using API key and secret:

curl -X 'GET' \
  'https://apis.threatwinds.com/api/ai/v1/models/silas-1.0' \
  -H 'accept: application/json' \
  -H 'api-key: your-api-key' \
  -H 'api-secret: your-api-secret'

Response

Success Response (200 OK)

{
  "id": "claude-sonnet-4",
  "object": "model",
  "name": "Claude Sonnet 4.5",
  "provider": "claude",
  "owned_by": "Anthropic PBC",
  "created": 2025,
  "capabilities": [
    "chat",
    "tools-use",
    "reasoning",
    "code-generation",
    "image"
  ],
  "limits": {
    "max_input_tokens": 64000,
    "max_completion_tokens": 64000,
    "max_total_tokens": 128000
  }
}

Response Schema

Field	Type	Description
id	string	Model unique identifier
object	string	Object type, always “model”
name	string	Human-readable model name
provider	string	Provider identifier
owned_by	string	Organization that owns the model
created	integer	Year of model release
capabilities	array	List of model capabilities
limits	object	Token limit information
limits.max_input_tokens	integer	Maximum input tokens accepted
limits.max_completion_tokens	integer	Maximum tokens model can generate
limits.max_total_tokens	integer	Maximum combined tokens

Business Logic

• Retrieves model by ID from provider registry
• Returns 404 if model ID doesn’t exist
• No database queries - data from provider configuration

Error Codes

Status Code	Description	Possible Cause
200	OK	Model found and returned
404	Not Found	Model not found
400	Bad Request	Invalid request
401	Unauthorized	Missing or invalid authentication
403	Forbidden	Insufficient permissions

---

Available Models

Claude Models (Anthropic)

Model ID	Name	Max Input	Max Output	Capabilities
claude-sonnet-4	Claude Sonnet 4.5	64,000	64,000	Chat, tools, reasoning, code-generation, image
claude-opus-4	Claude Opus 4.5	32,000	32,000	Chat, tools, reasoning, code-generation, image
claude-haiku-4	Claude Haiku 4.5	64,000	64,000	Chat, tools, image

Internal Mappings:

• claude-sonnet-4 → claude-sonnet-4-5-20250929
• claude-opus-4 → claude-opus-4-1-20250805
• claude-haiku-4 → claude-haiku-4-5

Best For:

• Sonnet 4.5: Balanced performance with extended reasoning support, largest context (64K)
• Opus 4.5: Highest quality reasoning and complex problem-solving (32K context)
• Haiku 4.5: Fast responses for simple tasks, cost-effective (no reasoning)

Groq Models

Model ID	Name	Max Input	Max Output	Capabilities
gpt-oss-20b	GPT OSS 20B	131,072	65,536	Chat, code, tools, reasoning
gpt-oss-120b	GPT OSS 120B	131,072	65,536	Chat, code, tools, reasoning
qwen3-32b	Qwen 3 32B	131,072	40,960	Chat, code, tools, reasoning*
llama4-maverick	LLaMA 4 Maverick 17B	131,072	8,192	Chat, code, tools
llama4-scout	LLaMA 4 Scout 17B	131,072	8,192	Chat, code, tools

Internal Mappings:

• gpt-oss-20b → openai/gpt-oss-20b
• gpt-oss-120b → openai/gpt-oss-120b
• qwen3-32b → qwen/qwen3-32b
• llama4-maverick → meta-llama/llama-4-maverick-17b-128e-instruct
• llama4-scout → meta-llama/llama-4-scout-17b-16e-instruct

Best For:

• Extremely fast inference (optimized hardware)
• Real-time applications
• Cost-effective at scale
• Large context windows (131K tokens)

Special Notes:

• *qwen3-32b: reasoning_effort is accepted but always treated as “default” internally
• Token counting not supported for Groq models

vLLM Models (ThreatWinds)

Model ID	Name	Max Input	Max Output	Capabilities
silas-1.0	Silas 1.0	100,000	65,536	Chat, tools, code-generation

Best For:

• Cybersecurity-focused tasks
• Threat intelligence analysis
• Penetration testing assistance
• Security operations

Special Notes:

• Silas is a specialized cybersecurity AI assistant developed by ThreatWinds
• Optimized for offensive security and threat intelligence tasks
• Token counting supported via vLLM tokenization

---

Choosing the Right Model

By Use Case

Use Case	Recommended Models
Simple Q&A	Claude Haiku, Groq models
Complex Reasoning	Claude Opus, Claude Sonnet
Code Generation	Claude Sonnet, Silas
Long Context	Groq models (131K), Claude models (64K)
Vision Tasks	Claude Sonnet, Claude Opus, Claude Haiku
Real-time Chat	Groq models, Claude Haiku
Tool/Function Calling	Claude Sonnet, Claude Opus
Cybersecurity	Silas, Claude Sonnet

By Performance Priority

Priority	Models
Speed	Groq (LLaMA, GPT OSS), Claude Haiku
Quality	Claude Opus, Claude Sonnet
Cost	Claude Haiku, Groq models
Context Length	Groq (131K), Claude (64K)
Cybersecurity Expertise	Silas

By Provider Strengths

Provider	Strengths
Claude (Anthropic)	Instruction following, safety, reasoning, vision
Groq	Ultra-fast inference, low latency, large context
vLLM (ThreatWinds)	Cybersecurity specialization, pentesting support

---

Example: Filtering Models

List Only Claude Models

curl -X 'GET' 'https://apis.threatwinds.com/api/ai/v1/models' \
  -H 'Authorization: Bearer <token>' | \
  jq '.data[] | select(.provider == "claude")'

Find Models with Vision

curl -X 'GET' 'https://apis.threatwinds.com/api/ai/v1/models' \
  -H 'Authorization: Bearer <token>' | \
  jq '.data[] | select(.capabilities | contains(["image"]))'

Find Models with Large Context

curl -X 'GET' 'https://apis.threatwinds.com/api/ai/v1/models' \
  -H 'Authorization: Bearer <token>' | \
  jq '.data[] | select(.limits.max_input_tokens > 100000)'

Find Models with Tool Support

curl -X 'GET' 'https://apis.threatwinds.com/api/ai/v1/models' \
  -H 'Authorization: Bearer <token>' | \
  jq '.data[] | select(.capabilities | contains(["tools-use"]))'

---

Model Selection Best Practices

1. Check Token Limits: Ensure model supports your required context length
2. Verify Capabilities: Confirm model has required features (vision, tools, etc.)
3. Consider Cost: Balance quality needs with token costs
4. Test Multiple Models: Compare results across models for your use case
5. Monitor Performance: Track response times and quality metrics
6. Stay Updated: Model list changes as providers release new versions
7. Match Expertise: Use specialized models (like Silas) for domain-specific tasks