One gateway. Every model.

Seamlessly route requests to the optimal AI model based on cost, latency, or capability. Unified API, zero friction.

GPT-4o

OpenAI

Operational

Context Window

128k

Max Output

16k

Vision
Tool Use
JSON Mode

GPT-4o-mini

OpenAI

Operational

Context Window

128k

Max Output

16k

Vision
Tool Use
JSON Mode

Claude 3.5 Sonnet

Anthropic

Operational

Context Window

200k

Max Output

8k

Vision
Tool Use
Computer Use

Claude 3 Haiku

Anthropic

Operational

Context Window

200k

Max Output

4k

Vision
Tool Use

Llama 3.1 70B

Together AI

Operational

Context Window

128k

Max Output

8k

Tool Use
JSON Mode

Llama 3.1 8B

Together AI

Operational

Context Window

128k

Max Output

4k

JSON Mode

Mixtral 8x7B

OpenRouter

Operational

Context Window

32k

Max Output

4k

JSON Mode

Gemini 1.5 Pro

Google

Operational

Context Window

1M

Max Output

8k

Vision
Tool Use
Audio

Llama 3.2 90B

Groq

Operational

Context Window

128k

Max Output

8k

Tool Use
JSON Mode
SAPI — One API. Every model. Zero overhead.