GPT 5 Nano

gpt-5-nano

GPT 5 Nano is available on Swarms Cloud through our Agent Completions API, no separate provider account or API key required. Run it as a standalone agent with the Agent Completions API, enable tools, structured outputs, and autonomous loops, or orchestrate it inside multi-agent swarms alongside models from other providers. GPT 5 Nano also supports vision tasks: pass base64-encoded images and it can analyze, describe, and answer questions about them.

Open in Playground

Loading catalog details…

Quick start

Get started with GPT 5 Nano

Get your Swarms API key

Create an API key from your API keys dashboard or at swarms.world/platform/api-keys.

Set it in your environment

# .env
SWARMS_API_KEY="your-api-key"

# or export it in your shell
export SWARMS_API_KEY="your-api-key"

Run a single agent

Execute one agent with gpt-5-nano via the Agent Completions API.

Single agent completion

POST /v1/agent/completions

curl -X POST 'https://api.swarms.world/v1/agent/completions' \
  -H 'x-api-key: '"$SWARMS_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
  "agent_config": {
    "agent_name": "Research Analyst",
    "description": "Expert in analyzing and synthesizing research data",
    "system_prompt": "You are a Research Analyst with expertise in data analysis and synthesis.",
    "model_name": "gpt-5-nano",
    "max_loops": 1,
    "max_tokens": 8192,
    "temperature": 0.5
  },
  "task": "Analyze the impact of artificial intelligence on healthcare"
}'

Scale to a multi-agent swarm

Chain multiple agents on gpt-5-nano with the Swarm Completions API.

Multi-agent swarm

POST /v1/swarm/completions

curl -X POST 'https://api.swarms.world/v1/swarm/completions' \
  -H 'x-api-key: '"$SWARMS_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Research Swarm",
  "description": "A two-agent research and analysis pipeline",
  "swarm_type": "SequentialWorkflow",
  "task": "Research the latest developments in renewable energy and summarize the top 3 trends",
  "agents": [
    {
      "agent_name": "Researcher",
      "description": "Gathers and organizes source material",
      "system_prompt": "You are a meticulous researcher. Collect relevant facts and organize them clearly.",
      "model_name": "gpt-5-nano",
      "max_loops": 1
    },
    {
      "agent_name": "Analyst",
      "description": "Synthesizes research into actionable insights",
      "system_prompt": "You are an analyst. Turn the research into a concise, actionable summary.",
      "model_name": "gpt-5-nano",
      "max_loops": 1
    }
  ],
  "max_loops": 1
}'

Grab these docs as Markdown

This page is also available as a plain Markdown file, no HTML parsing required — handy for pasting into an LLM or letting an agent fetch it directly.

curl -s https://cloud.swarms.world/models/gpt-5-nano.md

Frequently asked questions

How do I use GPT 5 Nano with the Swarms API?

Set your SWARMS_API_KEY environment variable, then send a POST request to /v1/agent/completions with agent_config.model_name set to "gpt-5-nano". The API works from Python, TypeScript, cURL, or any HTTP client.

Can I use GPT 5 Nano in a multi-agent swarm?

Yes. Pass "gpt-5-nano" as the model_name of any agent in a POST to /v1/swarm/completions. You can mix it with other models across 17+ swarm architectures such as SequentialWorkflow, ConcurrentWorkflow, and HierarchicalSwarm.

How is GPT 5 Nano billed on Swarms Cloud?

Usage is billed per input and output token. See the Swarms Cloud pricing page for the token cost calculator and current rates.

Do I need a separate provider API key to use GPT 5 Nano?

No. A single Swarms API key gives you access to every model in the catalog through one agent completions API, no separate provider accounts required.