Model Catalog

The LLM API supports 35+ models from 4 providers, managed dynamically via the llm_models catalog.

Providers

Provider	Main Models
OpenAI	GPT-4.1, GPT-4.1-mini, GPT-4.1-nano, o3, o3-mini, o4-mini
Anthropic	Claude Opus 4.6, Claude Sonnet 4.5, Claude Haiku 4.5
Google	Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash Lite
Grok	Grok 3, Grok 3 Mini, Grok 4 Fast, Grok Code Fast

Model ID Format

Models use the provider.model format:

openai.gpt-4.1-nano
anthropic.claude-sonnet-4-5-20250929
google.gemini-2.5-flash
grok.grok-3-mini

Use auto to let the system select the best model for your task (see Auto-Routing).

---

Model Tier Access by Plan

Models are grouped into tiers. Your subscription plan determines which tiers you can access:

Plan	Allowed Tiers	Example Models
free	economical	GPT-4.1-nano, Haiku 4.5, Gemini 2.5 Flash Lite
basic	economical, premium	+ GPT-4.1, Sonnet 4.5, Gemini 2.5 Flash
pro	economical, premium, flagship	+ Claude Opus 4.6, o3, Gemini 2.5 Pro
enterprise	all	+ legacy models

Requesting a model outside your plan's allowed tiers returns HTTP 403:

{
  "error": "MODEL_TIER_RESTRICTED",
  "message": "Model 'claude-opus-4-6' (tier: flagship) is not available for the 'basic' plan.",
  "details": {
    "model_tier": "flagship",
    "plan_code": "basic",
    "allowed_tiers": ["economical", "premium"]
  }
}

tip

Use model: "auto" to automatically select the best model within your plan's allowed tiers.

---

Feature Support by Model

Not all models support every feature. The tools_support field in the catalog tracks per-model capabilities:

Feature	Description	Unsupported Models
function_calling	Tool/function calling	None (all active models)
structured_output	JSON Schema response	Anthropic models
stop_sequences	Custom stop sequences	Reasoning models (see below)

Stop Sequences Compatibility

Reasoning models do not support the stop parameter. The API silently skips it for these models.

Provider	Models WITHOUT stop support	Reason
OpenAI	o1, o3, o3-mini, o4-mini, gpt-5, gpt-5-mini, gpt-5-nano	Reasoning models reject stop
Grok	grok-3-mini, grok-4-0709, grok-4-fast-reasoning, grok-4-1-fast-reasoning	xAI returns 400 for reasoning models
Anthropic	(none)	All models support stop
Google	(none)	All models support stop

Querying Stop Support

Use the tools_support.stop_sequences field from the model catalog to check programmatically:

SELECT model_id, tools_support->>'stop_sequences' as stop_support
FROM llm_models WHERE is_active = true;

---

GET /api/llm/models

List all available models.

Authentication: Not required

Cache: 5 minutes

Response:

{
  "success": true,
  "count": 38,
  "models": [
    {
      "id": "openai.o3-mini",
      "name": "o3-mini",
      "provider": "openai",
      "description": "Fast and cost-effective reasoning model",
      "tier": "premium",
      "context": "200000",
      "capabilities": ["reasoning", "code", "math"],
      "performance": {
        "latency": 7,
        "quality": 8,
        "cost": 7
      },
      "auto_routing_enabled": true
    }
  ]
}

Model Fields:

Field	Type	Description
id	string	Full model ID (provider.model)
name	string	Model name
provider	string	Provider name
description	string	Short description
tier	string	Model tier: economical, balanced, premium, flagship
context	string	Context window size (tokens)
capabilities	array	Model capabilities (reasoning, code, vision, etc.)
performance	object	Performance scores (1-10 scale)
auto_routing_enabled	boolean	Whether the model participates in auto-routing

---

GET /api/v3/llm/models/stats

Get global usage statistics per model.

Authentication: Not required

Cache: 15 minutes

Query Parameters:

Param	Type	Default	Description
days	integer	30	Number of days to analyze (1-365)

Response:

{
  "success": true,
  "period_days": 30,
  "generated_at": "2026-01-18T12:00:00.000Z",
  "data": [
    {
      "provider": "openai",
      "model": "gpt-4o",
      "total_tokens": 12500000,
      "total_requests": 5000,
      "avg_latency_ms": 450.25,
      "last_used": "2026-01-18T11:30:00.000Z"
    }
  ],
  "count": 15
}

---

GET /api/llm/provider/:provider

Get information about a specific provider.

Authentication: Not required

Cache: 15 minutes

Parameters:

Param	Type	Description
provider	string	openai, anthropic, google, grok

Response:

{
  "success": true,
  "provider": "openai",
  "supportedModels": [
    "gpt-4.1", "gpt-4.1-mini", "gpt-4.1-nano",
    "gpt-4o", "gpt-4o-mini", "o3", "o3-mini", "o4-mini"
  ],
  "modelCount": 8
}