Models

Playbooks is tested and optimized for the following models:

• Anthropic Claude Sonnet 4 ("claude-sonnet-4-20250514")
• PlaybooksLM

Playbooks framework uses complex prompts that are optimized for Sonnet 4. As a result, other frontier LLMs such as GPT-5, Gemini 2.5 Pro and Grok 4 may work, but additional testing and tuning may be required. Our current focus is on improving the core capabilities of the framework and demonstrating a novel way of programming AI agents. We plan to add testing and support for other models gradually.

As of Sept 2025, we recommend using Sonnet 4 for experimentation and development.

For production use, you may continue using Sonnet 4 or switch to PlaybooksLM for lower latency, lower cost and full control over data privacy & security. PlaybooksLM is a proprietary, Enterprise-grade model developed by Playbooks AI. Please contact us at contact@runplaybooks.ai for more information.