> ## Documentation Index
> Fetch the complete documentation index at: https://docs.together.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Overview
> Run inference on 100+ open-source models.
Together AI offers three ways to run inference:
**[Serverless models](/docs/serverless/models):** A shared fleet of popular open models you can call through a per-token API. No GPUs to provision or manage. Best for prototyping, or apps with variable traffic.
**[Provisioned throughput](/docs/inference/provisioned-throughput):** Reserved capacity for a selected stock model with a defined SLA covering committed throughput and reliability. Best for production workloads that need stronger guarantees than serverless.
**[Dedicated endpoints](/docs/dedicated-endpoints/overview):** A single model running on GPUs reserved for you, billed per minute by hardware. Best for serving fine-tuned models, or workloads that need direct control over hardware, latency, and throughput.
## Get started
Set up an API key and make your first call in Python, TypeScript, or cURL.
Our picks for common inference use cases.
How Together AI bills for inference.
## Shared inference API
Serverless, provisioned throughput, and dedicated endpoint workloads all use the same inference APIs for generating and retrieving model outputs. Apps work on any deployment mode without code changes; just swap the `model` parameter:
```python Python highlight={7,13} theme={null}
from together import Together
client = Together()
# Serverless model request
response = client.chat.completions.create(
model="moonshotai/Kimi-K2.6",
messages=[{"role": "user", "content": "Hello!"}],
)
# Dedicated endpoint request
response = client.chat.completions.create(
model="/Qwen/Qwen3.5-9B-FP8-bb04c904",
messages=[{"role": "user", "content": "Hello!"}],
)
```
```typescript TypeScript highlight={6,12} theme={null}
import Together from "together-ai";
const client = new Together();
// Serverless model request
let response = await client.chat.completions.create({
model: "moonshotai/Kimi-K2.6",
messages: [{ role: "user", content: "Hello!" }],
});
// Dedicated endpoint request
response = await client.chat.completions.create({
model: "/Qwen/Qwen3.5-9B-FP8-bb04c904",
messages: [{ role: "user", content: "Hello!" }],
});
```
```bash cURL highlight={6,15} theme={null}
# Serverless model request
curl -X POST "https://api.together.ai/v1/chat/completions" \
-H "Authorization: Bearer $TOGETHER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "moonshotai/Kimi-K2.6",
"messages": [{"role": "user", "content": "Hello!"}]
}'
# Dedicated endpoint request
curl -X POST "https://api.together.ai/v1/chat/completions" \
-H "Authorization: Bearer $TOGETHER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "/Qwen/Qwen3.5-9B-FP8-bb04c904",
"messages": [{"role": "user", "content": "Hello!"}]
}'
```
## Integrations
Drop-in replacement for OpenAI clients.
Together SDKs and framework wiring.
## Batch processing
If your workload doesn't need a real-time response, submit it as a batch job for up to 50% off serverless rates.
## Model capabilities
Chat completions, streaming, parameters.
Tool use and agentic loops.
Pass images alongside text.
FLUX, Kontext, and Google models.
Text-to-video and image-to-video.
Batch and streaming transcription.
HTTP and WebSocket audio output.
Vectors, rerankers, and RAG.