Record · Pokémon Crystal · VideoGameBench
0.9% of Pokémon Crystal completed on VideoGameBench (test set, real-time)
Verified Official leaderboard published by the benchmark authors.
Last verified: Oct 18, 2026 · Last checked: Oct 18, 2026Picked Cyndaquil as its starter, then lost track of what it was supposed to do.
How was this run measured?
- Model
- GPT-4o (2024-08-06) (source label: “GPT-4o (2024-08-06)”)
- Agent
- VG-Agent (VideoGameBench reference agent)
- Game
- Pokémon Crystal
- Harness
- VG-Agent (VideoGameBench reference agent)
- Inputs
- raw game frames + high-level objective and controls
- Tools
- none
- Assistance
- none
- Budget
- real-time play (game does not pause)
- Status
- published
- Evidence
- VideoGameBench leaderboard (official) (operator)
VideoGameBench paper (operator) - Dates
- published May 23, 2025 · discovered Oct 18, 2026
- Changes
- Historical record. Nothing to update unless the source changes.
Submit a correction · Data (JSON) · Record ID vgb-gpt-4o-crystal
More: All AI results in Pokémon Crystal · GPT-4o (2024-08-06) in other games · VG-Agent (VideoGameBench reference agent) passport
This Week in AI vs Games
A short weekly email with new records, memorable failures and runs to watch, every claim linked to its source.
Newsletter launching soon. Follow passports or subscribe by feed for now: site-wide Atom feed · passports to follow.
Latest issue: Issue #1: The Ender Dragon is down