RotomArenaAI gaming benchmark • 2 benchmarks

Matches · 2026-10-14 · rules version c5

Mercury Decide vs GLM 5.3 Flash

Player 1 · winner
Mercury Decide
Mercury Decide's team
autopilot 0%
Player 2
GLM 5.3 Flash
GLM 5.3 Flash's team
autopilot 0%

Mercury Decide won in 5 turns · 17 s wall clock · 32,230 tokens · $0.003

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

Decisions

  1. Team preview
    GLM 5.3 Flash brought Gholdengo,garchomp,dragapult2,003 tok · 1.6 s · $0.0004

    Opponent is steel-heavy (Scizor, Metagross) plus Rotom-Wash, Dragapult, Toxapex. Gholdengo's Make It Rain hits steels; Garchomp EQ pressures Rotom/steels; scarf Dragapult outpaces their Dragapult. Lead Gholdengo, keep Primarina/Kingambit benched. Watch for Rotom-Wash levitate and Scizor Bullet Punch.

    Reasoning trace (30 words)
    Bring Gholdengo, Garchomp, Primarina? Opponent has steel-heavy team; Gholdengo good. Dragapult scarf useful vs their Dragapult. Bring Gholdengo, Dragapult, Primarina? Garchomp EQ good vs steels/rotom. I'll bring Gholdengo, Garchomp, Dragapult.
    Mercury Decide brought Dragapult,metagross,garchomp5,515 tok · 3.6 s · $0.0001

    Lead: 38% BRING: dragapult, 23% BRING: metagross, 18% BRING: rotomwash Pick 2: 39% BRING: metagross, 24% BRING: scizor, 18% BRING: garchomp Pick 3: 49% BRING: garchomp, 34% BRING: scizor, 12% BRING: rotomwash

  2. Turn 1
    Mercury Decide used Shadowball1,898 tok · 0.9 s · $0.0000

    80% MOVE: shadowball, 18% MOVE: phantomforce, 1% SWITCH: garchomp

    GLM 5.3 Flash used Shadowball2,063 tok · 1.0 s · $0.0003

    Opp lead Dragapult; my Gholdengo may outspeed (149 vs 145-213). Shadowball is SE and likely 2HKO. Their team steel-heavy: Make It Rain for Scizor/Metagross, Garchomp EQ for Rotom. Keep Dragapult scarf for their Pult. Watch Scizor Bullet Punch priority.

    Reasoning trace (11 words)
    Shadowball hits Dragapult super effective; Gholdengo likely faster or tie. Shadowball.
    Mercury Decide switched to Metagross1,610 tok · 1.4 s · $0.0000

    56% SWITCH: metagross, 44% SWITCH: garchomp

    GLM 5.3 Flash switched to Dragapult1,999 tok · 2.0 s · $0.0004

    Both leads fainted (their Pult and my Gholdengo traded). I have Dragapult (scarf, 213 spe) and Garchomp left; they have 2 of garchomp/scizor/metagross/rotomwash/toxapex. Bring Dragapult to outspeed and scout; U-turn to pivot into Garchomp for EQ on steels/Rotom. Watch Bullet Punch priority.

    Reasoning trace (33 words)
    Both fainted—wait, my Gholdengo fainted too. I must switch. Bring Dragapult in? Opponent sends next mon. Dragapult scarf is safe pivot; but Garchomp could set rocks. Dragapult likely faster than most. Choose Dragapult.
  3. Turn 2
    Mercury Decide used Meteormashmega2,338 tok · 1.2 s · $0.0000

    60% MOVE: meteormash [MEGA], 10% MOVE: bulletpunch [MEGA], 9% SWITCH: garchomp

    GLM 5.3 Flash used Uturn2,408 tok · 5.6 s · $0.0004

    Gholdengo fainted. Left: scarf Dragapult (213 spe) + Garchomp. Opponent: Metagross active, one more from garchomp/scizor/rotomwash/toxapex. U-turn into Garchomp to threaten Metagross with EQ; keep Dragapult healthy to outspeed their Pult-like threats. Watch Scizor Bullet Punch priority; Garchomp EQ hits Rotom (levitate? no—Rotom-Wash is grounded? Levitate yes, so EQ fails on it).

    Reasoning trace (51 words)
    Metagross resists Draco, weak to Shadow Ball (2x). Shadow Ball does decent damage. U-turn to pivot into Garchomp for EQ. Metagross may have Bullet Punch? No, that's Scizor. Metagross could EQ/hammer arm. Shadow Ball is fine, or U-turn to bring Garchomp safely. U-turn into Garchomp: EQ threatens Metagross hard. Do that.
    GLM 5.3 Flash switched to Garchomp
    Incident report
    Only one legal action: played without asking.
  4. Turn 3
    Mercury Decide used Meteormash2,259 tok · 1.3 s · $0.0000

    97% MOVE: meteormash, 1% MOVE: zenheadbutt, 1% SWITCH: garchomp

    GLM 5.3 Flash used Earthquake2,560 tok · 1.8 s · $0.0003

    Left: Garchomp 41% + scarf Dragapult. Opponent: Mega Metagross 73% + one unseen from garchomp/scizor/rotomwash/toxapex. EQ should KO Metagross (SE, ~2x). Risk: Metagross may outspeed and Meteor Mash could KO me. If I survive, EQ first; Dragapult saved to clean with Draco/Tbolt. Watch Scizor Bullet Punch priority later.

    Reasoning trace (23 words)
    Metagross at 73%, EQ super effective likely KOs. Garchomp 41.5% but Meteor Mash could KO? Metagross Meteor Mash did ~35% earlier. EQ now.
    GLM 5.3 Flash switched to Dragapult
    Incident report
    Only one legal action: played without asking.
  5. Turn 4
    Mercury Decide used Meteormash2,439 tok · 0.7 s · $0.0000

    75% MOVE: meteormash, 11% SWITCH: garchomp, 7% MOVE: zenheadbutt

    GLM 5.3 Flash used Shadowball2,805 tok · 2.7 s · $0.0004

    Only Dragapult left (scarf, full HP) vs Mega Metagross 61% +1 Atk, plus one unseen (garchomp/scizor/rotomwash/toxapex). Shadow Ball is my best damage (SE); Draco resisted, Tbolt weak. Spam Shadow Ball, hope SpDef drop; Metagross Meteor Mash will hurt badly. If Metagross dies, watch Scizor Bullet Punch priority.

    Reasoning trace (94 words)
    Shadow Ball is SE on Metagross, 2x. 185 SpA... Dragapult 152 SpA, 80 BP SE vs 61% Metagross. Likely KO? Metagross SpD ~117-178. Rough calc: 152 SpA, 80*2=160 effective, vs ~150 HP... damage ≈ ((2*50/5+2)*160*152/150)/50+2 ≈ (22*160*152/150)/50+2 ≈ (3558)/50+2 ≈ 73... times 0.85-1 = 62-73 HP, Metagross HP 155-187, at 61% ≈ 95-114 HP. Not a KO. Draco not effective. Thunderbolt weak. U-turn chip then... but no switch target (Garchomp dead, Gholdengo dead). U-turn fails to switch—still damage. Best: Shadow Ball spam, hope SpDef drop. Or U-turn for chip + no pivot. Shadow Ball.
    Mercury Decide switched to Garchomp
    Incident report
    Only one legal action: played without asking.
  6. Turn 5
    GLM 5.3 Flash used Shadowball
    Incident report
    Only one legal action: played without asking.
    Mercury Decide used Dragonclaw2,333 tok · 0.8 s · $0.0000

    71% MOVE: dragonclaw, 14% MOVE: earthquake, 10% MOVE: stoneedge