Skip to main content

Models

FRONTIER MODELS

300+ frontier LLM models from OpenAI, Anthropic, Google, and more.

Overview

All frontier models are accessible via POST /v1/chat/completions using the model's full ID (e.g. openai/gpt-5.4).

Model IDProviderContext
openai/gpt-5.4OpenAI1M
openai/gpt-5.4-miniOpenAI1M
anthropic/claude-sonnet-4.6Anthropic1M
anthropic/claude-opus-4.6Anthropic1M
google/gemini-3.1-pro-previewGoogle1M
google/gemini-3-flash-previewGoogle1M
google/gemini-3.1-flash-liteGoogle1M
x-ai/grok-4.20xAI256K
qwen/qwen3.5-plusQwen128K
moonshotai/kimi-k2Moonshot128K
openrouter/autoAuto Router

Auto Router

openrouter/auto uses NotDiamond to automatically select the best model for your prompt:

python
response = client.chat.completions.create(
    model="openrouter/auto",
    messages=[{"role": "user", "content": "Your prompt here"}]
)

Provider Routing

Control which upstream serves your request via the provider parameter (raw HTTP body):

json
{
  "model": "openai/gpt-5.4",
  "messages": [...],
  "provider": {
    "sort": "price",
    "order": ["OpenAI", "Azure"],
    "only": ["OpenAI"],
    "ignore": ["Azure"],
    "max_price": {"input": 5, "output": 15}
  }
}

> OpenAI Python SDK rejects unknown kwargs (TypeError: Completions.create() got an unexpected keyword argument 'provider'). Pass these fields under extra_body instead:

>

> ```python

> client.chat.completions.create(

> model="openai/gpt-5.4",

> messages=[...],

> extra_body={"provider": {"sort": "throughput", "order": ["OpenAI", "Azure"]}},

> )

> ```

FieldValuesDescription
sort"price""latency""throughput"Optimization priority
orderarray of provider namesPreferred provider order
onlyarrayRestrict to these providers only
ignorearrayExclude these providers
max_price{input, output}Max price per 1M tokens

Plugins

Extend model capabilities with plugins:

json
{
  "plugins": [
    {"id": "web"},
    {"id": "file-parser"},
    {"id": "response-healing"},
    {"id": "context-compression"}
  ]
}
PluginDescription
webReal-time web search
file-parserParse PDFs and documents
response-healingAuto-repair malformed JSON responses
context-compressionCompress long contexts to fit within limits

Model Fallbacks

Specify fallback models if the primary is unavailable:

json
{
  "model": "anthropic/claude-opus-4.6",
  "models": ["anthropic/claude-opus-4.6", "openai/gpt-5.4"],
  "route": "fallback"
}

Model Shortcuts

Append suffixes to any model ID for routing hints:

SuffixDescription
:nitroThroughput priority — fastest response
:floorLowest price available

Example: openai/gpt-5.4:nitro, anthropic/claude-sonnet-4.6:floor

Was this page helpful?