> ## 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.
# Choosing the Right Model for Your Portrix Workload
> Compare speed, cost, and capability across 400+ models on Portrix. Learn how to pick the right model for chat, reasoning, embeddings, and vision tasks.
Choosing the right model is about balancing cost, latency, capability, and context window size for your specific workload. With 400+ models available on Portrix, the options can feel overwhelming — this guide cuts through the noise so you can make a confident decision and start iterating fast.
## Model ID format
Every model on Portrix is identified by a `{provider}/{model-name}` string. You pass this string as the `model` parameter in any request:
```
openai/gpt-4o
anthropic/claude-3-5-sonnet
google/gemini-2.0-flash
mistral/mistral-large
```
The provider prefix routes the request to the correct backend automatically — no additional configuration required.
## Specifying the model via header
If your HTTP client or framework does not support modifying the request body, you can pass the model ID in the `x-portrix-model` header instead of the `model` field. The header takes precedence when both are present:
```bash theme={null}
curl https://api.portrix.ai/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-H "x-portrix-model: anthropic/claude-3-5-sonnet" \
-d '{"messages": [{"role": "user", "content": "Hello!"}]}'
```
This is particularly useful when proxying requests through middleware that cannot rewrite the body, or when building multi-tenant systems where the calling layer selects the model on behalf of the application.
## By use case
These models excel at general-purpose conversation, instruction following, and multi-step reasoning tasks.
| Model | Strengths |
| ----------------------------- | ---------------------------------------------------------------- |
| `openai/gpt-4o` | Versatile, fast, strong reasoning and instruction following |
| `anthropic/claude-3-5-sonnet` | Nuanced writing, long-form reasoning, low hallucination rate |
| `google/gemini-2.0-flash` | Very low latency, cost-efficient for high-volume workloads |
| `mistral/mistral-large` | Strong multilingual support, competitive European data residency |
```python theme={null}
response = client.chat.completions.create(
model="openai/gpt-4o",
messages=[{"role": "user", "content": "Explain transformer attention in simple terms."}],
)
```
Use these models when your input — documents, codebases, transcripts — exceeds typical context limits.
| Model | Context Window |
| ----------------------------- | ---------------- |
| `anthropic/claude-3-5-sonnet` | 200,000 tokens |
| `google/gemini-1.5-pro` | 1,000,000 tokens |
```python theme={null}
with open("long_document.txt") as f:
document = f.read()
response = client.chat.completions.create(
model="google/gemini-1.5-pro",
messages=[
{"role": "user", "content": f"Summarise this document:\n\n{document}"}
],
)
```
Embeddings convert text into dense vectors for semantic search, clustering, and retrieval-augmented generation (RAG).
| Model | Dimensions | Best For |
| ------------------------------- | ---------- | ---------------------------------------- |
| `openai/text-embedding-3-small` | 1,536 | Cost-efficient, high-throughput indexing |
| `openai/text-embedding-3-large` | 3,072 | Highest accuracy retrieval tasks |
```python theme={null}
response = client.embeddings.create(
model="openai/text-embedding-3-small",
input="Portrix is a unified AI gateway.",
)
vector = response.data[0].embedding
```
Vision models accept image inputs alongside text, enabling document understanding, screenshot analysis, and multimodal reasoning.
| Model | Notes |
| ------------------------- | --------------------------------------------------------- |
| `openai/gpt-4o` | Strong OCR, diagram interpretation, and image Q\&A |
| `google/gemini-2.0-flash` | Fast and cost-efficient for image classification at scale |
```python theme={null}
response = client.chat.completions.create(
model="openai/gpt-4o",
messages=[
{
"role": "user",
"content": [
{"type": "text", "text": "What is in this image?"},
{"type": "image_url", "image_url": {"url": "https://example.com/chart.png"}},
],
}
],
)
```
These models perform strongly on code generation, debugging, and technical explanation tasks.
| Model | Strengths |
| ----------------------------- | ------------------------------------------------------------ |
| `openai/o3-mini` | Deep reasoning for hard algorithmic and engineering problems |
| `anthropic/claude-3-5-sonnet` | Clean, well-commented code with excellent test generation |
```python theme={null}
response = client.chat.completions.create(
model="anthropic/claude-3-5-sonnet",
messages=[
{"role": "user", "content": "Write a Python function to binary-search a sorted list."}
],
)
```
## Cost vs quality tradeoffs
Use this table as a starting point — actual pricing is available in your [Portrix dashboard](https://app.portrix.ai).
| Model | Provider | Relative Cost | Relative Speed | Best For |
| ------------------------------- | --------- | ------------- | -------------- | -------------------------------------- |
| `google/gemini-2.0-flash` | Google | \$ | ⚡⚡⚡ | High-volume, latency-sensitive apps |
| `mistral/mistral-large` | Mistral | \$ | ⚡⚡⚡ | Multilingual and EU-resident workloads |
| `openai/gpt-4o` | OpenAI | \$\$ | ⚡⚡ | General-purpose production workloads |
| `google/gemini-1.5-pro` | Google | \$\$ | ⚡⚡ | Very long document processing |
| `anthropic/claude-3-5-sonnet` | Anthropic | \$\$ | ⚡⚡ | Writing, reasoning, low hallucination |
| `openai/o3-mini` | OpenAI | \$\$\$ | ⚡ | Complex coding and math reasoning |
| `openai/text-embedding-3-small` | OpenAI | \$ | ⚡⚡⚡ | Embedding at scale |
| `openai/text-embedding-3-large` | OpenAI | \$\$ | ⚡⚡ | High-precision semantic retrieval |
## Listing all available models
Retrieve the full, live list of models available to your account at any time:
```python Python theme={null}
models = client.models.list()
for model in models.data:
print(model.id)
```
```bash cURL theme={null}
curl https://api.portrix.ai/v1/models \
-H "Authorization: Bearer YOUR_API_KEY"
```
The response is a standard OpenAI-compatible `ModelList` object. Each entry includes the model ID, provider, and any capability flags your account has access to.
## Using model routing
If you want Portrix to select the best model automatically based on a routing strategy, set the `x-portrix-route` header instead of specifying a model directly:
```bash theme={null}
curl https://api.portrix.ai/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "x-portrix-route: cheapest" \
-H "Content-Type: application/json" \
-d '{"messages": [{"role": "user", "content": "Hello!"}]}'
```
Supported routing strategies include `cheapest`, `fastest`, and `balanced`. See the Fallbacks & Load Balancing guide for the full routing reference.
Start with a cheaper, faster model like `google/gemini-2.0-flash` during development and evaluation. Once you've validated your prompts and logic, benchmark against a more capable model to decide whether the quality improvement justifies the cost increase.