Examples demonstrating how to run Swarms agents against different model providers. Each subfolder is a provider (or model family) and contains runnable, self-contained scripts.
Set the matching API key as an environment variable before running any example (e.g. OPENAI_API_KEY, ANTHROPIC_API_KEY, OPENROUTER_API_KEY, AZURE_API_KEY, GROQ_API_KEY).
- claude_4.py — Claude 4 agent
- claude_4_example.py — Claude 4 integration
- swarms_claude_example.py — Swarms + Claude
- fable_agent.py — Claude Fable 5 agent
- fable_example.py — Claude Fable 5 example
- azure_agent.py — Azure OpenAI integration
- azure_agent_api_verison.py — Azure API version handling
- azure_model_support.py — Azure model support
- cerebras_example.py — Cerebras model integration
- deepseek_r1.py — DeepSeek R1 (native API)
- fast_r1_groq.py — DeepSeek R1 served via Groq
- groq_deepseek_agent.py — DeepSeek on Groq agent
- gpt_os_agent.py — GPT-OSS agent
- groq_gpt_oss_models.py — GPT-OSS on Groq
- multi_agent_gpt_oss_example.py — Multi-agent GPT-OSS
- llama_4.py — Llama 4 agent
- litellm_example.py — Llama 4 via LiteLLM
- simple_agent.py — Simple Llama 4 agent
- lumo_example.py — Lumo model integration
- mistral_example.py — Mistral integration
- simple_example_ollama.py — Local models via Ollama
- gpt_5_example.py — GPT-5 integration
- 5.4_openai.py — GPT-5.4 via Swarms client
- 4o_mini_demo.py — GPT-4o mini demo
- concurrent_gpt5.py — Concurrent GPT-5 processing
- example_o3.py — o3 reasoning example
- o3_agent.py — o3 reasoning agent
- reasoning_duo_batched.py — Batched reasoning duo
- test_async_litellm.py — Async LiteLLM test
- qwen_3_base.py — Qwen 3 base model
- vllm_example.py — vLLM integration
- vllm_wrapper.py — vLLM wrapper
- base_llm.py — Plug a custom LLM class into an
Agent