// Games
Solve today's Wordle
Play the NYT Wordle end to end: read the tile colors after every guess and pick the next word until it is solved.
16 agent tool calls, click to jump
Why it’s hard
A game is a loop of acting, reading state and deciding, where every guess depends on the last board.
- Tile colors only appear after a flip animation settles
- Guesses outside Wordle's word list are rejected and must be retried
- Duplicate letters follow Wordle's own scoring rules
Stagehand vs. Playwright MCP
Given the same goal and model, the Stagehand code-mode agent finished for 32% less than an agent driving Playwright MCP.
- stagehand code mode agent3/3 runs succeeded
- $0.180
- playwright mcp agent3/3 runs succeeded
- $0.263
medians over successful runs, or over all runs when none succeeded; lower is better
Then run it as a script
Once the flow works, the cookbook runs it as plain Stagehand calls. With caching on, a re-run cost 19% more than the first run. 1 of 3 cached re-runs failed, so treat that figure with care.
- cookbook, first run2/3 runs succeeded
- $0.035
- cookbook, cached re-run2/3 runs succeeded
- $0.042
medians over successful runs, or over all runs when none succeeded; lower is better
anthropic/claude-sonnet-5 at $2 input · $0.2 cached · $2.5 cache write · $10 output per 1m tokens · stagehand 4.1.0 · playwright mcp 0.0.82 · agents stop after 60 llm calls · run on 2026-09-28 · methodology
Run it yourself
Give the same goal to Claude Code with Stagehand’s code-mode MCP server, which exposes three tools: run, snapshot and screenshot. Once the flow works, the cookbook runs it as a plain Stagehand script.
Play today's Wordle at nytimes.com/games/wordle and solve it in six guesses or fewer. Report whether you solved it, the answer (null if not solved), and each guess in order with its tile feedback as five emoji (🟩 correct, 🟨 present, ⬛ absent).