History
What changed, and when
Two timelines: the automatic change log from our collectors, and every record by the date its evidence was published. If nothing moved, this page says nothing new. We don't invent news.
Change log
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turns 8,710 → 8,825 · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 4 Argon Turns 3,707 → 3,722 · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turns 7,898 → 8,710 reviewed · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Badges 4 → 5 reviewed · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 4 Argon Turns 3,518 → 3,707 reviewed · source
- New record (reviewed): Claude Opus 5 kills the Ender Dragon in a fresh Minecraft world in one unbroken 3h 56m session reviewed · source
- New record (reviewed): GPT-6 Astra + Jev kill the Ender Dragon in 8m 43s on a fixed seed with a pre-surveyed route (Peaceful) reviewed · source
- New record (reviewed): Jev beats Pokémon Red in 37h 40m with a guided, open-source harness reviewed · source
- New record (reviewed): Jev enters the Pokémon Red Hall of Fame with Claude Opus 5 rebuilding its options live reviewed · source
- New record (reviewed): Gemini 2.5 Pro beats Pokémon Blue with no manual interventions (~406 hours, 36,801 actions) reviewed · source
- New record (reviewed): Gemini 2.5 Pro becomes Champion in Pokémon Blue (development run, ~813 hours) reviewed · source
- KradleVerse: AI agents play Minecraft: Claude Opus 4.6 30-day win % 24% → 25% · source
- VideoGameBench (real-time): now tracking Gemini 2.0 Flash (Overall score: 0%) reviewed · source
- VideoGameBench (real-time): now tracking Llama 4 Maverick (17B-128E Instruct FP8) (Overall score: 0%) reviewed · source
- VideoGameBench (real-time): now tracking Claude 3.7 Sonnet (2025-02-19) (Overall score: 0%) reviewed · source
- VideoGameBench Lite (game pauses while the model thinks): now tracking Gemini 2.0 Flash (Overall score: 0%) reviewed · source
- VideoGameBench Lite (game pauses while the model thinks): now tracking Llama 4 Maverick (17B-128E Instruct FP8) (Overall score: 0%) reviewed · source
- VideoGameBench Lite (game pauses while the model thinks): now tracking Gemini 2.5 Pro (preview 03-25) (Overall score: 1.6%) reviewed · source
- VideoGameBench Lite (game pauses while the model thinks): now tracking Claude 3.7 Sonnet (2025-02-19) (Overall score: 1.6%) reviewed · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turns 7,890 → 7,898 · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 4 Argon Turns 3,509 → 3,518 · source
- KradleVerse: AI agents play Minecraft: Claude Opus 4.6 30-day win % 23% → 24% · source
- Kaggle Game Arena: Werewolf: now tracked automatically
- Kaggle Game Arena: Werewolf: status RECENT → ARCHIVE
- Kaggle Game Arena: Heads-up Poker: now tracked automatically
- Kaggle Game Arena: Go: now tracked automatically
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Badges 3 → 4 · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): now tracking Gemini 4 Argon (Badges: 8) · source
- Gemini Flash 'vision-first' series: Pokémon Blue: status UNVERIFIED → RECENT
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turn 7,177 → 7,178 · source
- ClaudePlaysPokemon: Red (Claude Opus 4.7): status LIVE → PAUSED
- ClaudePlaysPokemon: Red (Claude Opus 4.7): Claude Opus 4.7 Live now yes → no · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turn 7,175 → 7,177 · source
- Gemini Flash 'vision-first' series: Pokémon Blue: status PAUSED → UNVERIFIED
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turn 7,167 → 7,175 · source
- Gemini Flash 'vision-first' series: Pokémon Blue: now tracked automatically
- Gemini Flash 'vision-first' series: Pokémon Blue: status UNVERIFIED → PAUSED
- ClaudePlaysPokemon: Red (Claude Opus 4.7): status PAUSED → LIVE
- ClaudePlaysPokemon: Red (Claude Opus 4.7): Claude Opus 4.7 Live now no → yes · source
- Pokémon VGC Bench: now tracked automatically
- Pokémon Autonomous Completion (BenchmarkList): now tracked automatically
- KradleVerse: AI agents play Minecraft: now tracked automatically
- KradleVerse: AI agents play Minecraft: Claude Opus 4.6 30-day win % 22% → 23% · source
- Kaggle Game Arena: Chess (text): now tracked automatically
- GPT Plays Pokémon (vision-only harness): now tracked automatically
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): now tracked automatically
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Location Lavaridge Town Pokémon Center 1F → Lavaridge Town Gym B1F · source
- Gemini Plays Pokémon: Emerald (Gemini 4 Argon & Gemini 3.8 Flash): Gemini 3.8 Flash Turn 7,109 → 7,167 · source
- Factorio Learning Environment: now tracked automatically
- ClaudePlaysPokemon: Red (Claude Opus 4.7): now tracked automatically
- BALROG: BabyAI, Crafter, TextWorld, BabaIsAI, MiniHack, NetHack: now tracked automatically
- BalatroBench: now tracked automatically
- BalatroBench: status UNVERIFIED → ARCHIVE
Records by published date
- published Oct 18, 2026 · discovered Oct 18, 2026
Pokémon Emerald · Gemini 4 Argon: All 8 badges in Pokémon Emerald (now 3,722 turns in, still playing) Verified - published Oct 18, 2026 · discovered Oct 18, 2026
Pokémon Emerald · Gemini 4 Argon: Entered the Hall of Fame in Pokémon Emerald (beat Champion Wallace) Creator-reported - published Oct 18, 2026 · discovered Oct 17, 2026
Pokémon Emerald · Gemini 3.8 Flash: 5 of 8 badges in Pokémon Emerald after 8,825 turns (live) Verified - published Oct 18, 2026 · discovered Oct 17, 2026
Minecraft · Claude Opus 4.6: #1 agent on KradleVerse: 25% win rate over the last 30 days Verified - published Oct 15, 2026 · discovered Oct 18, 2026
Minecraft · Gemini 4 Argon: A team of 4 Gemini 4 Argon agents beat Minecraft: killed the Ender Dragon in 27h 14m, with 605 deaths and 2,783 programs written Creator-reported - published Oct 15, 2026 · discovered Oct 17, 2026
Chess · Gemini 3.8 Flash: #1 on Kaggle Game Arena Chess: 1452 Elo Verified - published Oct 15, 2026 · discovered Oct 18, 2026
Go · GPT-6 Astra: #1 on Kaggle Game Arena Go: 1794 Elo Verified - published Oct 14, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 3.8 Flash: 3 of 8 badges in Pokémon Blue, run finished after 5,648 turns Verified - published Oct 13, 2026 · discovered Oct 18, 2026
Pokémon (undisclosed game / ROM hack) · Gemini 3.8 Flash: 8 badges on run 'sors' (38,783 turns) Awaiting verification - published Oct 11, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 4 Argon: 8 of 8 badges in Pokémon Blue, run finished after 1,753 turns Verified - published Oct 8, 2026 · discovered Oct 18, 2026
Minecraft · GPT-6 Astra: Top Mine Odyssey score: 85.6% of 180 navigation tasks in Minecraft rebuilds of real places Creator-reported - published Oct 8, 2026 · discovered Oct 18, 2026
Heads-up Poker · GPT-5.6 Sol: #1 on Kaggle Game Arena Heads-up Poker: 30 Mean BB/100 Verified - published Sep 28, 2026 · discovered Oct 18, 2026
Pokémon Red · Jev: Jev beats Pokémon Red in 37h 40m with a guided, open-source harness Creator-reported - published Sep 27, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 3.8 Flash: 8 of 8 badges in Pokémon Blue (47,889 turns, idle) Verified - published Sep 27, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 3.8 Flash: Became Champion of Pokémon Blue at turn 24,981 (Hall of Fame at turn 24,994) Creator-reported - published Sep 27, 2026 · discovered Oct 18, 2026
Pokémon Red · Jev: Jev enters the Pokémon Red Hall of Fame with Claude Opus 5 rebuilding its options live Creator-reported - published Sep 20, 2026 · discovered Oct 18, 2026
Minecraft · GPT-6 Astra: GPT-6 Astra + Jev kill the Ender Dragon in 8m 43s on a fixed seed with a pre-surveyed route (Peaceful) Creator-reported - published Sep 19, 2026 · discovered Oct 17, 2026
Pokémon Blue · Gemini 3.7 Flash: 8 of 8 badges in Pokémon Blue (55,601 turns, idle) Verified - published Sep 19, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 3.7 Flash: Became Champion of Pokémon Blue at turn 39,310 Creator-reported - published Sep 18, 2026 · discovered Oct 17, 2026
NetHack · GPT-6 Astra: 13.2% progress on NetHack for the top overall BALROG entry. Nobody has beaten the game. Verified - published Sep 16, 2026 · discovered Oct 18, 2026
Minecraft · Claude Opus 5: Claude Opus 5 kills the Ender Dragon in a fresh Minecraft world in one unbroken 3h 56m session Creator-reported - published Sep 5, 2026 · discovered Oct 17, 2026
Pokémon (game not stated) · GPT-6 Astra: Became Champion in 18h 12m, the fastest claimed AI Champion time we know of Creator-reported - published Sep 3, 2026 · discovered Oct 17, 2026
Pokémon Blue · Gemini 3.5 Flash: 6 of 8 badges in Pokémon Blue (74,636 turns, idle) Verified - published Aug 15, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 3.6 Flash: 4 of 8 badges in Pokémon Blue at turn ~55,800 (last repo commit Aug 15, 2026) Creator-reported - published Jul 21, 2026 · discovered Oct 17, 2026
Pokémon Emerald · GPT-5.6 Sol: Became Champion of Pokémon Emerald in 153h51 (8 badges) Awaiting verification - published Jun 28, 2026 · discovered Oct 18, 2026
Minecraft · Claude Opus 4.6: Top MineExplorer score: 41.08 overall task success (open-world exploration) Creator-reported - published Jun 12, 2026 · discovered Oct 18, 2026
Werewolf · GPT-5.4: #1 on Kaggle Game Arena Werewolf: 0.08% Equilibrium Rating Verified - published May 23, 2026 · discovered Oct 18, 2026
Pokémon Blue · Gemini 3 Pro: All 8 badges in a vision-only Pokémon Blue run (last turn 67,315, May 2026) Creator-reported - published Mar 16, 2026 · discovered Oct 18, 2026
Pokémon Emerald · Heatz agent (model not disclosed): 1st: reached the first gym (Roxanne) in 40:13 (1,608 actions) Verified - published Mar 16, 2026 · discovered Oct 18, 2026
Pokémon Emerald · Hamburg PokéRunners agent (model not disclosed): 2nd: reached the first gym in 1:14:43 Verified - published Mar 16, 2026 · discovered Oct 18, 2026
Pokémon Emerald · anthonys agent (model not disclosed): 3rd: reached the first gym in 1:29:17 Verified - published Mar 16, 2026 · discovered Oct 18, 2026
Pokémon Emerald · Deepest agent (model not disclosed): Judge's Choice: fewest actions to the first gym (649), 5th by time Verified - published Dec 15, 2025 · discovered Oct 18, 2026
Pokémon Crystal · Gemini 3: Reached the post-game with 16 badges in Pokémon Crystal by turn 36,290 (race vs Gemini 2.5) Creator-reported - published Dec 15, 2025 · discovered Oct 18, 2026
Pokémon Crystal · Gemini 2.5: Crystal race opponent of Gemini 3; ran to turn 51,280 (Dec 2025). Final progress not disclosed. Awaiting verification - published Aug 25, 2025 · discovered Oct 17, 2026
Pokémon Crystal · GPT-5: Became Champion of Pokémon Crystal in 202h02m (9,517 steps, 16 badges) Creator-reported - published Aug 14, 2025 · discovered Oct 17, 2026
Pokémon Red · GPT-5: Became Champion of Pokémon Red in 161h57m (6,470 steps, 8 badges) Creator-reported - published Jun 9, 2025 · discovered Oct 18, 2026
Pokémon Blue · Gemini 2.5 Pro: Gemini 2.5 Pro beats Pokémon Blue with no manual interventions (~406 hours, 36,801 actions) Creator-reported - published May 23, 2025 · discovered Oct 18, 2026
Pokémon Crystal · GPT-4o (2024-08-06): 0.9% of Pokémon Crystal completed on VideoGameBench (test set, real-time) Verified - published May 23, 2025 · discovered Oct 18, 2026
Kirby's Dream Land · Gemini 2.5 Pro (preview 03-25): 4.8% of Kirby's Dream Land (best single-game score on the VideoGameBench test set; 0.48% overall) Verified - published May 2, 2025 · discovered Oct 18, 2026
Pokémon Blue · Gemini 2.5 Pro: Gemini 2.5 Pro becomes Champion in Pokémon Blue (development run, ~813 hours) Creator-reported
This Week in AI vs Games
One short email every Sunday: new records, memorable failures and runs to watch. Free. Every claim is linked to its source.
Latest issue: Issue #1: The Ender Dragon is down