# Let an AI model play and finish Pokemon Blue live on a stream

Software engineer Joel Zhang built a harness that let Google's Gemini 2.5 Pro play Pokemon Blue on a Twitch stream. The model finished the game after about 37 days, which Google CEO Sundar Pichai shared publicly.

- Blaze Score: 60/100
- Proof: Reported by several people
- Difficulty: Builder
- Tools: Gemini
- Source: https://www.cryptopolitan.com/googles-gemini-just-beat-pokemon-blue/
- URL: https://www.blazefish.com/use-case/let-an-ai-model-play-and-finish-pokemon-blue-live-on-a-stream/

### Steps

1. Run a Game Boy emulator you can control from code
2. Send the model a screenshot plus key game state (location, party) on each turn
3. Give it a small set of tools: press buttons, and save and read notes
4. Log its reasoning and stream the run so you can see where it gets stuck
5. Tweak the harness, not the game, when it loops

### Prompt

```
Write a Python harness that connects a Game Boy emulator (PyBoy) to the Gemini API. Each turn, capture a screenshot and read the player's map location from memory, send both to the model with a short goal, and let it reply with a JSON list of button presses. Keep a notes file the model can append to and read back. Log every turn to a file.
```