> ## Documentation Index
> Fetch the complete documentation index at: https://docs.portix.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Portrix Model Catalog: 400+ AI Models from Every Provider

> Portrix provides access to 400+ models from OpenAI, Anthropic, Google, Mistral, and more. Learn how models are identified and how to choose the right one.

Portrix gives you unified access to more than 400 models from many AI providers, all addressable through a single API. Whether you want OpenAI's GPT-4o, Anthropic's Claude, Google's Gemini, or a Meta Llama model running on Groq, you reach every one of them through the same endpoint using the same request format.

## Model naming convention

Portrix identifies every model using a `{provider}/{model-name}` format. This convention tells the gateway which provider to route your request to and which model to invoke. You pass this identifier as the `model` field in your request body, just as you would with any OpenAI-compatible API.

| Model ID                      | Provider               | Notes                        |
| ----------------------------- | ---------------------- | ---------------------------- |
| `openai/gpt-4o`               | OpenAI                 | Latest multimodal flagship   |
| `openai/gpt-4o-mini`          | OpenAI                 | Fast, cost-efficient variant |
| `anthropic/claude-3-5-sonnet` | Anthropic              | Strong reasoning and coding  |
| `anthropic/claude-3-haiku`    | Anthropic              | Lightweight, low-latency     |
| `google/gemini-2.0-flash`     | Google                 | Fast Gemini model            |
| `google/gemini-1.5-pro`       | Google                 | Long-context model           |
| `mistral/mistral-large`       | Mistral                | Mistral's flagship model     |
| `mistral/mistral-7b-instruct` | Mistral                | Open-weight efficient model  |
| `meta/llama-3.1-70b-instruct` | Meta (via Groq/others) | Open-weight Llama 3.1        |

## Supported providers

Portrix aggregates models from a wide range of providers. New providers and models are added continuously.

<CardGroup cols={3}>
  <Card title="OpenAI" icon="o">
    GPT-4o, GPT-4o mini, GPT-4 Turbo, o1, o3, text-embedding models, and more.
  </Card>

  <Card title="Anthropic" icon="a">
    Claude 3.5 Sonnet, Claude 3.5 Haiku, Claude 3 Opus, and the full Claude 3 family.
  </Card>

  <Card title="Google" icon="g">
    Gemini 2.0 Flash, Gemini 1.5 Pro, Gemini 1.5 Flash, and text embedding models.
  </Card>

  <Card title="Mistral" icon="m">
    Mistral Large, Mistral Small, Mixtral 8x7B, Codestral, and open-weight variants.
  </Card>

  <Card title="Meta (Llama)" icon="l">
    Llama 3.1 and Llama 3.2 models in various sizes, served through multiple infrastructure providers.
  </Card>

  <Card title="Cohere" icon="c">
    Command R+, Command R, and Embed models for retrieval-augmented generation.
  </Card>

  <Card title="Groq" icon="bolt">
    Ultra-low-latency inference for Llama, Mixtral, and Gemma models.
  </Card>

  <Card title="And more" icon="ellipsis">
    Perplexity, Together AI, Fireworks, DeepSeek, Qwen, and dozens of additional providers.
  </Card>
</CardGroup>

## Model capabilities

Different models support different capabilities. Not every model handles vision input, tool calls, or embeddings — check the model's metadata before building features that depend on specific capabilities.

| Capability | Description                           | Example models                                     |
| ---------- | ------------------------------------- | -------------------------------------------------- |
| Chat       | Multi-turn conversational completions | `openai/gpt-4o`, `anthropic/claude-3-5-sonnet`     |
| Completion | Single-turn text generation (legacy)  | `openai/gpt-3.5-turbo-instruct`                    |
| Embeddings | Dense vector representations of text  | `openai/text-embedding-3-large`, `cohere/embed-v3` |
| Vision     | Image understanding alongside text    | `openai/gpt-4o`, `google/gemini-1.5-pro`           |
| Tool use   | Structured function / tool calling    | `openai/gpt-4o`, `anthropic/claude-3-5-sonnet`     |
| Streaming  | Token-by-token SSE responses          | Supported by most chat models                      |

To see all available models and their capabilities programmatically, call the `/v1/models` endpoint:

```bash theme={null}
curl https://api.portrix.ai/v1/models \
  -H "Authorization: Bearer $PORTRIX_API_KEY"
```

<Note>
  Use `GET /v1/models` to programmatically list all available models and their metadata, including supported capabilities, context window sizes, and pricing information.
</Note>

You can also list models using the OpenAI SDK pointed at the Portrix base URL:

```python theme={null}
from openai import OpenAI

client = OpenAI(
    api_key="your-portrix-api-key",
    base_url="https://api.portrix.ai/v1",
)

models = client.models.list()

for model in models.data:
    print(model.id)
```

## Choosing a model

Selecting the right model involves balancing several factors. Here is a quick guide to get you started.

**Cost vs. quality** — Flagship models like `openai/gpt-4o` and `anthropic/claude-3-5-sonnet` deliver the highest quality but are more expensive per token. Smaller models like `openai/gpt-4o-mini` or `anthropic/claude-3-haiku` are significantly cheaper and still capable for many tasks.

**Latency** — If your use case is interactive (for example, a customer-facing chatbot), prioritize models with low time-to-first-token. Groq-hosted models and mini/flash variants are optimized for speed.

**Context window** — Long-document analysis, large codebases, or multi-turn conversations with extensive history require a large context window. Models like `google/gemini-1.5-pro` (1M tokens) and `anthropic/claude-3-5-sonnet` (200K tokens) are strong choices here.

**Capability requirements** — If you need vision input or tool calling, verify the model supports those features using the `/v1/models` endpoint before committing to it in production.

For a comprehensive comparison of models across these dimensions, see the [Model Selection guide](/guides/model-selection).