LLM Playground

Reverse Wordle

Pick a model and a 5-letter target word, then watch the LLM try to solve it in six guesses.

Model

Moonshot AI
Qwen
DeepSeek
Google
MiniMax
OpenAI
Z.AI
Anthropic
xAI

Solver: Claude Sonnet 4.6

Target word

Pick a 5-letter word for the model to solve. The model only sees its own guesses and the green/yellow/gray feedback after each one.

Prompt & game state sent to the LLM

Every turn the LLM is sent the same system prompt plus a user message with all prior guesses and their green/yellow/gray feedback. Click Show to inspect or edit.