Prompting advice changes - so we keep checking it.
Most prompt libraries are frozen the day someone published them. The techniques that helped in 2020 2023 2025 - elaborate chain-of-thought scaffolding, heavy few-shot examples, "think step by step" - can now hurt output on current reasoning models. This site tracks what's actually current, model by model, and shows its work.
Start here
Model guides
Current guidance for ChatGPT, Claude, Gemini, Perplexity, DeepSeek, Gemma, Grok, Llama, and Mistral.
Task-based recipes
Curated starting points for common tasks, rewritten against current guidance and tagged with which models they're validated against.
Prompting concepts
Evergreen explanations for system prompts, context windows, structured output, tools, examples, and reasoning models.
Model comparisons
Decision pages for choosing between model families, context options, cost tiers, and hosted or open-weight deployment paths.
Prompt Pattern Checker
Paste a prompt and review older prompting patterns before they become reusable advice.
Glossary
Plain-language definitions for the prompting and model terms used across this site's guides and recipes.
Tokens and context windows
Why the whole conversation resends every turn, what context rot is, and why figuring out token cost is hard.