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 ID | Provider | Context |
|---|---|---|
openai/gpt-5.4 | OpenAI | 1M |
openai/gpt-5.4-mini | OpenAI | 1M |
anthropic/claude-sonnet-4.6 | Anthropic | 1M |
anthropic/claude-opus-4.6 | Anthropic | 1M |
google/gemini-3.1-pro-preview | 1M | |
google/gemini-3-flash-preview | 1M | |
google/gemini-3.1-flash-lite | 1M | |
x-ai/grok-4.20 | xAI | 256K |
qwen/qwen3.5-plus | Qwen | 128K |
moonshotai/kimi-k2 | Moonshot | 128K |
openrouter/auto | Auto Router | — |
Auto Router
openrouter/auto uses NotDiamond to automatically select the best model for your prompt:
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):
{
"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"]}},
> )
> ```
| Field | Values | Description | ||
|---|---|---|---|---|
sort | "price" | "latency" | "throughput" | Optimization priority |
order | array of provider names | Preferred provider order | ||
only | array | Restrict to these providers only | ||
ignore | array | Exclude these providers | ||
max_price | {input, output} | Max price per 1M tokens |
Plugins
Extend model capabilities with plugins:
{
"plugins": [
{"id": "web"},
{"id": "file-parser"},
{"id": "response-healing"},
{"id": "context-compression"}
]
}| Plugin | Description |
|---|---|
web | Real-time web search |
file-parser | Parse PDFs and documents |
response-healing | Auto-repair malformed JSON responses |
context-compression | Compress long contexts to fit within limits |
Model Fallbacks
Specify fallback models if the primary is unavailable:
{
"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:
| Suffix | Description |
|---|---|
:nitro | Throughput priority — fastest response |
:floor | Lowest price available |
Example: openai/gpt-5.4:nitro, anthropic/claude-sonnet-4.6:floor