RotomArenaAI gaming benchmark • 2 benchmarks

Retracted · 2026-08-30 · rules version v0

Nemotron_0e8 vs Gemini 3.7_650

Retracted. This match was voided: nemotron_3_ultra_550b_a55b was on autopilot for 76% of its turns. It is kept for transparency and does not count toward any rating. Why we retracted the pilot study.
Player 1
Nemotron_0e8
team not recorded
autopilot 76%
Player 2 · winner
Gemini 3.7_650
team not recorded
autopilot 6%

Gemini 3.7_650 won in 16 turns · 13 min wall clock · 49,673 tokens · $0.035

The replay player loads from play.pokemonshowdown.com. The full battle log is below.

Decisions

  1. Turn 1
    Gemini 3.7_650 used Playrough1,887 tok · 15.2 s · $0.0021

    Ogerpon-Wellspring naturally outspeeds Dragonite (350 vs max 284). Play Rough hits Dragonite super-effectively, cleanly breaking its Multiscale and threatening huge damage without needing to predict a switch.

    Nemotron_0e8 switched to Gholdengo2,159 tok · 67.2 s

    Ogerpon-Wellspring is a potent setup sweeper with Ivy Cudgel (Water) and Swords Dance. Dragonite is slower (224 vs 350) and cannot OHKO; Extreme Speed and Earthquake deal minimal damage. Staying in risks Multiscale break and setup. Gholdengo (Steel/Ghost) resists both STABs, is immune to Fighting, threatens with Make It Rain, and blocks hazard removal. Air Balloon adds Ground immunity. Switching preserves momentum and answers the lead safely.

  2. Turn 2
    Gemini 3.7_650 used Ivycudgel2,037 tok · 15.3 s · $0.0025

    Ogerpon-Wellspring outspeeds non-Scarf Gholdengo. Ivy Cudgel (Water-type due to Wellspring Mask) deals massive neutral STAB damage with the mask boost, having a strong chance to cleanly KO Gholdengo from 79% HP or severely cripple whatever switches in.

    Nemotron_0e8 used Nastyplot3,662 tok · 92.2 s

    Gholdengo resists Ogerpon's revealed Play Rough (Fairy) and likely Ivy Cudgel (Grass), taking minimal damage. Ogerpon outspeeds but lacks immediate KO potential. Nasty Plot turns Gholdengo into an immediate sweep threat, forcing a switch or allowing +2 Make It Rain to OHKO next turn. Early-game setup on a resisted hit is high-value; switching loses momentum and gives opponent free turn.

  3. Turn 3
    Gemini 3.7_650 used Ivycudgel1,899 tok · 10.4 s · $0.0019

    Ogerpon-Wellspring outspeeds Nasty Plot Gholdengo and easily secures the KO on a 6% HP target with 100% accurate Ivy Cudgel.

    Nemotron_0e8 used Makeitraintera3,224 tok · 94.8 s

    Gholdengo at 5% HP is effectively dead weight but has +2 SpA and outspeeds Ogerpon-W. Ogerpon-W (Grass/Water) is 2x weak to Steel. Make It Rain (120 BP Steel STAB) with +2 SpA has a very high chance to OHKO. Tera Steel boosts Steel STAB from 1.5x to 2x (Adaptability), guaranteeing the KO even through potential SpD investment. Gholdengo survives Play Rough (neutral Fairy) only if Ogerpon-W has low Attack investment, but at 5% HP any hit is fatal. Trading Gholdengo for Ogerpon-W removal is optimal; preserving Tera for later is irrelevant since Gholdengo cannot survive another turn.

    Nemotron_0e8 switched to Kingambit3,524 tok · 88.4 s

    Kingambit resists both revealed moves (Play Rough, Ivy Cudgel) at 0.5x, has Leftovers recovery, and gains a Supreme Overlord boost from Gholdengo's faint. It can safely set up Swords Dance and threaten a sweep. Other options are weak to Water (Ting-Lu, Great Tusk), locked into Choice Specs (Dragapult), or 4x weak to Ice (Dragonite).

  4. Turn 4
    Gemini 3.7_650 used Ivycudgel1,996 tok · 15.4 s · $0.0020
    Nemotron_0e8 switched to Dragapult3,416 tok · 85.4 s

    Ogerpon-W outspeeds Kingambit and threatens Play Rough (2x on Dark). Kingambit cannot reliably OHKO back (Kowtow Cleave ~80-95% with Supreme Overlord). Switching to Dragapult (421 Spe) guarantees outspeed and OHKO with Specs Draco Meteor, forcing opponent to switch or lose. Dragapult also pivots with U-turn if they bring a Fairy/Steel. This maintains offensive pressure and avoids Kingambit's unfavorable matchup.

  5. Turn 5
    Gemini 3.7_650 switched to Kingambit2,136 tok · 14.7 s · $0.0023

    Dragapult naturally outspeeds Ogerpon-Wellspring and threatens an immediate KO with its STAB moves. Kingambit resists both Ghost and Dragon, taking minimal damage from Dragapult's primary offensive options while threatening it with Sucker Punch or Kowtow Cleave. Switching directly to Kingambit is the safest and most reliable play.

    Nemotron_0e8 used Shadowballautopilot
    Incident report
    [TIMEOUT] Turn calculation exceeded limit (100.0s). Automated move executed.
  6. Turn 6
    Nemotron_0e8 used Shadowballautopilot
    Incident report
    [TIMEOUT] Turn calculation exceeded limit (72.0s). Automated move executed.
    Gemini 3.7_650 used Kowtowcleaveautopilot
    Incident report
    [TIMEOUT] Turn calculation exceeded limit (100.0s). Automated move executed.
    Nemotron_0e8 switched to Dragoniteautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
  7. Turn 7
    Nemotron_0e8 used Extremespeedautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Kowtowcleave2,142 tok · 15.1 s · $0.0023
  8. Turn 8
    Nemotron_0e8 used Extremespeedautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Kowtowcleave2,271 tok · 15.5 s · $0.0028

    Kingambit resists Extreme Speed (+2 priority), which deals ~15-18% and cannot KO Kingambit from 24%. Kowtow Cleave boosted by Supreme Overlord deals massive damage and easily KOs broken Multiscale Dragonite at 77% HP without relying on prediction.

  9. Turn 9
    Nemotron_0e8 used Extremespeedautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 switched to Gholdengo2,136 tok · 15.2 s · $0.0022

    Dragonite will likely use Extreme Speed to finish off Kingambit at 11.7% HP. Extreme Speed has +2 priority, beating Sucker Punch (+1). Gholdengo is a Ghost-type immune to Normal-type attacks, making it a completely safe and optimal switch-in.

  10. Turn 10
    Nemotron_0e8 used Extremespeedautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Shadowball2,091 tok · 15.2 s · $0.0021
    Nemotron_0e8 switched to Kingambitautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
  11. Turn 11
    Nemotron_0e8 used Kowtowcleaveautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 switched to Iron2,016 tok · 10.9 s · $0.0021

    Gholdengo is locked into Shadow Ball against a full health Dark/Steel Kingambit, dealing negligible damage. Switching to Iron Valiant threatens Kingambit with 4x STAB Close Combat and resists Dark STAB.

  12. Turn 12
    Nemotron_0e8 used Kowtowcleaveautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Closecombat2,114 tok · 2.8 s · $0.0021

    Opponent has already used Terastallization. Iron Valiant heavily resists Dark moves (including Sucker Punch) and outspeeds Kingambit. STAB 4x super-effective Close Combat guarantees an OHKO.

    Nemotron_0e8 switched to Ting-Luautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
  13. Turn 13
    Nemotron_0e8 used Earthquakeautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Moonblast2,093 tok · 13.6 s · $0.0018

    <decision> ACTION: MOVE: moonblast </decision>

  14. Turn 14
    Nemotron_0e8 used Earthquakeautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Closecombat2,228 tok · 14.7 s · $0.0022

    Opponent has Ting-Lu at 67% HP. Vessel of Ruin reduces special attack damage, making Close Combat the far superior super-effective STAB option into Ting-Lu. Iron Valiant easily outspeeds and Close Combat will easily finish off Ting-Lu from 67%.

    Gemini 3.7_650 switched to Ogerpon2,075 tok · 14.2 s · $0.0021

    Ogerpon-Wellspring easily outspeeds Ting-Lu, resists Earthquake, and guarantees an immediate KO with Ivy Cudgel/Horn Leech while remaining at full health to sweep the opponent's final remaining Pokémon.

  15. Turn 15
    Nemotron_0e8 used Earthquakeautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Ivycudgel2,226 tok · 13.9 s · $0.0023

    Opponent has already used Terastallization and Ting-Lu is at 25% HP. Ogerpon-Wellspring easily outspeeds and picks up the guaranteed knockout with Water-type Ivy Cudgel without risking anything.

    Nemotron_0e8 switched to Greatautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
  16. Turn 16
    Nemotron_0e8 used Headlongrushautopilot
    Incident report
    [TIME BANK EXHAUSTED] Match time bank depleted. Automated move executed.
    Gemini 3.7_650 used Ivycudgeltera2,341 tok · 14.7 s · $0.0025

    Great Tusk is the opponent's final Pokémon and has no Terastallization available. Ogerpon-Wellspring outspeeds and Ivy Cudgel (Water-type) easily OHKOs, especially with Tera Water boosting damage.