Matches · 2026-10-11 · rules version c4

GLM 5.3 Flash vs Claude Haiku 5.5

Claude Haiku 5.5 won in 6 turns · 59 s wall clock · 41,382 tokens · $0.007

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

Decisions

1. Team preview
   GLM 5.3 Flash brought Garchomp,sylveon,gholdengo1,763 tok · 1.2 s · $0.0001

   Opponent has Sinistcha and Archaludon; my Sylveon handles both well, and Gholdengo blocks status and pressures them. Garchomp with sash leads to set rocks and scout. Kingambit covers Gholdengo matchups.

   Claude Haiku 5.5 brought Incineroar,gholdengo,kingambit3,465 tok · 6.1 s · $0.0008

   Opponent's Garchomp, Dragonite, and Sylveon are the main threats. Incineroar gives Intimidate and Parting Shot support, Gholdengo is a fast special attacker with Steel and Ghost coverage, and Kingambit handles Fairy and Dragon with Iron Head and strong physical pressure.

   Reasoning trace (75 words)

   Looking at team preview, I'm weighing a solid core of Incineroar for intimidate and pivoting support alongside Gholdengo's special offense, then deciding between Kingambit or Archaludon as the third pick against their Garchomp, Kingambit, Gholdengo, Rotom-Wash, Sylveon, and Dragonite.
   
   Kingambit's Iron Head is super effective against Sylveon since Steel hits Fairy hard, and it handles Garchomp fine too. I'll settle on Incineroar, Gholdengo, and Kingambit as my team, covering the main threats reasonably well.
   
   

2. Turn 1
   GLM 5.3 Flash used Earthquake1,319 tok · 1.1 s · $0.0001

   Incineroar is slower and Earthquake hits it super effective for heavy damage, likely a big chunk or KO. Attacking now is better than setting rocks while Intimidate already cut my Attack.

   Claude Haiku 5.5 used Fakeout2,290 tok · 2.9 s · $0.0003

   Garchomp is faster than Incineroar, so it would hit first. Fake Out has +3 priority, so it moves first and flinches Garchomp, which stops its attack this turn. Garchomp is already at -1 Attack from Intimidate, which helps too.

   Reasoning trace (37 words)

   Fake Out looks like the best opening move here since Incineroar is faster, it'll flinch Garchomp and prevent it from acting while also chipping neutral damage, on top of the Intimidate drop already weakening its attack.
   
   

3. Turn 2
   GLM 5.3 Flash used Stealthrock1,411 tok · 1.1 s · $0.0001

   Incineroar will likely Fake Out again or pivot; Earthquake hits it hard but Intimidate cuts my damage and it may just switch. Setting Stealth Rock now while I'm faster and safe pressures its whole team (Garchomp, Archaludon weak). Rocks is the best long-term play.

   Claude Haiku 5.5 used Partingshot3,153 tok · 7.5 s · $0.0008

   Incineroar is slower than Garchomp and Flare Blitz is resisted by Dragon-type Garchomp, so attacking is poor. Parting Shot lowers Garchomp's Attack and Special Attack and pivots Incineroar out. Kingambit can then take on Garchomp, since Steel resists Dragon moves.

   Reasoning trace (100 words)

   Garchomp's already weakened and its attack is lowered from intimidate, so I'm weighing whether to stay in with Incineroar or pivot out with Parting Shot to bring in Kingambit or Gholdengo safely.
   
   Flare Blitz is weak vs Garchomp, so that's a poor option, and Protect just stalls without real benefit. Parting Shot seems better, dropping Garchomp's Atk further while bringing in Kingambit, though it takes Earthquake damage on entry; Kingambit still looks like the stronger breaker overall thanks to Supreme Overlord.
   
   I'll commit to Parting Shot as the move -- the switch-in choice will likely come up separately afterward.
   
   

   Claude Haiku 5.5 switched to Gholdengo2,870 tok · 5.9 s · $0.0007

   Garchomp is at -2 Attack and -1 Special Attack, so its Earthquake hits much weaker. Kingambit's contact moves would take Rough Skin damage, and it is slower than Garchomp. Gholdengo is a non-contact special attacker with Make It Rain and Shadow Ball, and it's faster than Garchomp's likely range at 136.

   Reasoning trace (71 words)

   Since I can only switch, I'm weighing Gholdengo versus Kingambit against Incineroar, checking type interactions against Garchomp's Dragon/Ground typing and Gholdengo's Steel/Ghost resistances and weaknesses, including whether Ground hits it super effectively.
   
   Gholdengo's special attacking means Garchomp's SpA drop is irrelevant, while Garchomp's own Atk drop blunts its Earthquake. Gholdengo likely outspeeds and definitely beats the slow Kingambit, so despite Life Orb recoil, Gholdengo seems like the right pick here.
   
   

4. Turn 3
   GLM 5.3 Flash used Earthquake1,605 tok · 2.0 s · $0.0001

   Gholdengo is Steel/Ghost, so Earthquake hits it super effectively and my Garchomp outspeeds, likely scoring a KO before it moves. My attack drops don't matter much since EQ is strong anyway.

   Reasoning trace (20 words)

   Gholdengo immune to EQ? No—Gholdengo is Steel/Ghost, Ground hits Steel, Ghost doesn't matter. EQ super effective, likely KO. Garchomp faster.

   Claude Haiku 5.5 used Makeitrain3,191 tok · 7.2 s · $0.0007

   Garchomp is at 85% with its Attack lowered by 2, so its physical damage is weak. Make It Rain (120 power, STAB, Life Orb) should roughly knock it out before it can do much, even with the Sp. Atk drop. Thunderbolt does nothing to a Ground type.

   Reasoning trace (101 words)

   Garchomp's at 85% with lowered attack and special attack, facing a nearly full-health Gholdengo. I'm checking type matchups: Shadow Ball is neutral against Garchomp, and Steel-type moves like Make It Rain aren't resisted by Dragon/Ground either, so Gholdengo's attacks hit normally here.
   
   Running the damage numbers, Make It Rain's 120 base power with STAB and Life Orb could get close to a KO on Garchomp despite the harsh SpA drop, while Shadow Ball only lands around 70%. Given that Garchomp's lowered attack already limits its offensive options, I think Make It Rain for the likely knockout is the better call.
   
   

   GLM 5.3 Flash switched to Gholdengo1,565 tok · 1.7 s · $0.0001

   Their Gholdengo is at 29% with -2 SpA after Life Orb recoil. My Gholdengo likely outspeeds (149 max vs theirs) and Make It Rain should KO it before it attacks. Switching in Gholdengo is the best play.

   Reasoning trace (44 words)

   Opponent Gholdengo at 29% with -2 SpA. My Gholdengo outspeeds likely (149 max vs theirs 93-149; mine is 149). Make It Rain from -2 won't kill mine; mine can KO theirs? Mine Make It Rain vs 29% Gholdengo — likely KO. Switch in Gholdengo.

5. Turn 4
   GLM 5.3 Flash used Shadowball1,657 tok · 2.0 s · $0.0001

   Opposing Gholdengo is at 29% and Shadow Ball is super effective, so it should knock out before Make It Rain can threaten me. Even in a speed tie, attacking first is worth the risk.

   Reasoning trace (14 words)

   Shadowball likely KOs (29% HP, SE, Life Orb). Speed tie risk but worth it.

   Claude Haiku 5.5 switched to Kingambit4,515 tok · 10.9 s · $0.0013

   Our Gholdengo is at 48 HP with -2 Sp. Atk, so Shadow Ball would not KO the full-HP opposing Gholdengo, and Life Orb recoil would leave it fragile. Kingambit takes only half damage from Ghost and Steel attacks, and Kowtow Cleave hits Gholdengo super effectively next turn.

   Reasoning trace (203 words)

   My Gholdengo is at -2 SpA, weakening its damage output, while the opponent's Gholdengo is at full HP as a Steel/Ghost type. Shadow Ball would be super effective against it since Ghost hits Ghost super effectively, so I'm weighing whether that 80 base power move can still get through despite my lowered stat.
   
   My life orb recoil is also a concern since I'm already low at 48 HP, and using Make It Rain would drop my SpA even further. I'm now considering whether switching to Kingambit, which is at full health, might be the safer play to absorb a hit instead.
   
   Checking Thunderbolt against Kingambit's Dark/Steel typing, Electric comes out neutral since Steel doesn't resist it. Then I'm weighing Sucker Punch's Dark-type damage against Gholdengo's Steel/Ghost typing, which takes neutral damage from Dark since Ghost isn't weak to it.
   
   Kingambit looks like the strongest switch-in since it'll tank a hit and threaten heavy damage back, but I'm also weighing just attacking now with Shadow Ball given my current SpA drop and Life Orb boost, roughly estimating the damage output against Gholdengo's HP and SpD.
   
   Stealth Rock just takes a mild 12.5% chip off Kingambit, nothing dangerous. I'll go with switching to Kingambit.
   
   

6. Turn 5
   GLM 5.3 Flash used Nastyplot1,810 tok · 1.5 s · $0.0001

   Kingambit resists both my attacks, so attacking now does little. It's much slower and can't hit me super effectively, so Nasty Plot is safe setup—though Sucker Punch risk exists, boosting lets me threaten the rest of their team.

   Reasoning trace (43 words)

   Kingambit resists both attacks; Nasty Plot while it likely attacks? Kingambit could Sucker Punch. Nasty Plot risky but Make It Rain drops our SpA. Shadow Ball is weak. Nasty Plot is best—Kingambit can't hit Gholdengo super effectively (Sucker Punch neutral, Iron Head neutral).

   Claude Haiku 5.5 used Kowtowcleave3,891 tok · 8.7 s · $0.0010

   Kingambit's Kowtow Cleave is Dark, which is super effective on Gholdengo's Ghost typing, and with Black Glasses and STAB it should knock out the 90% Gholdengo in one hit. Sucker Punch is weaker and fails if Gholdengo sets up with Nasty Plot, so Kowtow Cleave is the stronger play.

   Reasoning trace (160 words)

   My Kingambit at 106/177 HP is facing Gholdengo at 90%, which is Steel/Ghost and immune to status thanks to Good as Gold. Since Dark is super effective against Ghost but neutral against Steel, Kowtow Cleave with STAB and Black Glasses should still hit decently hard.
   
   Running the damage math, Kowtow Cleave looks like it could deal around 150-200+ damage, potentially enough to knock out Gholdengo depending on its defense roll, while Sucker Punch would be noticeably weaker at base 70 power and risky since it only works if Gholdengo is attacking that turn. Kowtow Cleave seems like the stronger, more reliable choice here.
   
   But I need to consider speed: if Gholdengo outspeeds Kingambit and uses Nasty Plot, Sucker Punch fails entirely, while Kowtow Cleave still lands. There's also the risk that Gholdengo could hit Kingambit back with Make It Rain, though its Steel typing means it's not very effective against Kingambit's Steel/Dark typing.
   
   Kowtow Cleave is the safer choice.
   
   

   GLM 5.3 Flash switched to Sylveon1,603 tok · 1.1 s · $0.0001

   Gholdengo fainted; only Sylveon remains. Sylveon's Hyper Voice (Pixilate) hits Kingambit neutrally but Kingambit outspeeds and likely KOs with Kowtow Cleave. Still, no other choice—bring Sylveon in.

7. Turn 6
   GLM 5.3 Flash used Yawn1,728 tok · 1.3 s · $0.0001

   Hyper Voice is resisted by Kingambit and Sylveon may be slower, so attacking is weak. Yawn pressures Kingambit to switch or risk sleep, letting Sylveon stall with Leftovers. Best play.

   Reasoning trace (17 words)

   Kingambit resists Hyper Voice (Dark/Steel). Sylveon slower possibly. Yawn forces Kingambit out or sleep. Yawn is good.

   Claude Haiku 5.5 used Ironhead3,546 tok · 6.5 s · $0.0008

   Kingambit is faster than Sylveon, and Iron Head is super effective on Fairy, so it should do most or all of Sylveon's HP in one hit. Kowtow Cleave is resisted by Fairy, and Sword Dance wastes a turn when a likely KO is available.

   Reasoning trace (87 words)

   I'm weighing Kingambit's matchup against Sylveon: Iron Head hits hard since Steel is super effective on Fairy, while Kowtow Cleave underperforms since Dark is resisted. I'm also factoring in Sylveon's potential Pixilate ability and its notably high special bulk and HP range.
   
   Running the numbers even with max defensive investment, Iron Head still deals roughly 170+ damage against Sylveon's likely 200ish max HP — not a guaranteed KO but very close, plus there's a flinch chance to consider. Iron Head is clearly the right call here.