AI plays Pokémon Blue: Gemini runs and records
Can AI beat Pokémon Blue? Gemini models have reached Champion. Turn counts, badge totals, harness details and sources for every AI Pokémon Blue run we track.
11 records · 1 runs or leaderboards
Pick a game to see every sourced AI result for it. Each game is scored on its own terms and never ranked against another game.
Can AI beat Pokémon Blue? Gemini models have reached Champion. Turn counts, badge totals, harness details and sources for every AI Pokémon Blue run we track.
11 records · 1 runs or leaderboards
Can AI beat Pokémon Emerald? Live Gemini runs, GPT results and PokéAgent Challenge speedruns, with verification labels and sources.
8 records · 2 runs or leaderboards
Can AI beat Pokémon Crystal? GPT-5 and Gemini results, plus how models do in real time on VideoGameBench. Sources and labels for every claim.
4 records · 0 runs or leaderboards
Can AI beat Pokémon Red? Sourced results for GPT-5, Claude and other models in Pokémon Red, with harness details and verification labels.
3 records · 1 runs or leaderboards
Yes, with caveats. Every sourced AI Pokémon run: Claude, Gemini and GPT in Red, Blue, Emerald and Crystal, with harness details and verification labels.
0 records · 2 runs or leaderboards
AI Pokémon results where the source doesn't name the exact game (a Pokémon game the source doesn't name). Kept separate so they are never mixed with Red…
1 records · 0 runs or leaderboards
AI Pokémon results where the source doesn't name the exact game (an undisclosed game or ROM hack). Kept separate so they are never mixed with Red, Blue…
1 records · 0 runs or leaderboards
Can AI play competitive Pokémon? Ratings of AI models in VGC doubles battles from Pokémon VGC Bench, with sources and verification labels.
0 records · 1 runs or leaderboards
Which AI is best at chess? Kaggle Game Arena's chess ladder for language models: Gemini 3.8 Flash is #1 as of 2026-10-15. Sourced and dated.
1 records · 1 runs or leaderboards
Which AI is best at Go? Kaggle Game Arena's Go ladder for language models: GPT-6 Astra is #1 as of 2026-10-15. Sourced and dated.
1 records · 1 runs or leaderboards
Which AI is best at heads-up poker? Kaggle Game Arena's heads-up poker ladder for language models: GPT-5.6 Sol is #1 as of 2026-10-08. Sourced and dated.
1 records · 1 runs or leaderboards
Which AI is best at Werewolf? Kaggle Game Arena's Werewolf ladder for language models: GPT-5.4 is #1 as of 2026-06-12. Sourced and dated.
1 records · 1 runs or leaderboards
Can AI beat Minecraft? A team of Gemini 4 Argon agents killed the Ender Dragon. Sourced Minecraft AI results, KradleVerse rankings and benchmarks.
6 records · 1 runs or leaderboards
Has AI beaten NetHack? Not yet. BALROG NetHack progress for the top AI models, with sources and dates.
1 records · 1 runs or leaderboards
VideoGameBench tests whether vision-language models can finish classic games from raw frames. Real-time and Lite results, kept separate.
0 records · 2 runs or leaderboards
Can AI play Age of Empires II? Language models wrote AoE2 AI scripts and played each other. What happened, with sources.
0 records · 1 runs or leaderboards
How good is AI at Balatro? BalatroBench ranks language models by the average round they reach. Scores, sources and data freshness.
0 records · 1 runs or leaderboards
Can AI play Factorio? In the Factorio Learning Environment, LLM agents write code to build factories. Production scores, sources and data freshness.
0 records · 1 runs or leaderboards
How far do AI models get in Kirby's Dream Land from raw frames? VideoGameBench results, with sources.
1 records · 0 runs or leaderboards
Orak (KRAFTON AI) tests language-model agents across 12 video games. Results and sources, never merged with other benchmarks.
0 records · 1 runs or leaderboards