ai benchmarks

Real-world AI performance metrics

Gauntlet Composite updated

Gauntlet Science updated

Gauntlet Ship updated

Gauntlet Teach updated

Gauntlet Judgment updated

Artificial Analysis data updated

Every change Artificial Analysis made to its scores and prices, newest first. Click a model to see its past scores.

1 price change

Price changes 1

  • single settingCost per task$0.29 → $0.31+$0.025

1 new release

New releases 1

  • peak 3.8one setting, $0.35 per task

1 new release

New releases 1

  • peak 64.65 effort levels, $3.10–$26.54 per task

1 new release

New releases 1

  • peak 36.02 effort levels, $0.12–$0.14 per task

1 new release · 1 price change

New releases 1

  • peak 24.0one setting, $0.57 per task

Price changes 1

  • highCost per task$0.16 → $0.17+$0.0085

1 price change

Price changes 1

  • single settingCost per task$0.54 → $0.52−$0.019

1 new release · 1 new effort level

New releases 1

  • peak 30.6one setting, $1.18 per task

New effort levels 1

  • highIntelligence40.5$4.56 per task

2 new releases

New releases 2

  • peak 44.2one setting, $11.82 per task
  • peak 34.6one setting, $1.03 per task

1 price change

Price changes 1

  • highCost per task$7.68 → $9.95+$2.26

1 new release

New releases 1

  • peak 23.8one setting, $0.33 per task

1 new release

New releases 1

  • peak 4.6one setting, $1.04 per task

1 new release

New releases 1

  • peak 7.4one setting, $0.53 per task

2 new releases · 1 score change

New releases 2

  • peak 31.0one setting, $4.68 per task
  • peak 6.2one setting, $0.19 per task

Score changes 1

  • maxIntelligence44.1 → 44.2+0.1

1 score change

Score changes 1

  • maxIntelligence32.2 → 32.3+0.1

1 new effort level

New effort levels 1

  • highIntelligence28.6$0.46 per task

51 score changes

Score changes 51

  • lowIntelligence34.5 → 34.8+0.3
  • Claude Opus 5.5mediumIntelligence48.1 → 48.2+0.1
  • Claude Opus 5.5highIntelligence51.7 → 51.8+0.1
  • Claude Opus 5.5xhighIntelligence57.0 → 57.4+0.4
  • Claude Opus 5.5maxIntelligence58.4 → 59.1+0.7
  • lowIntelligence53.7 → 54.2+0.5
  • Gemini 4 CarbonmediumIntelligence57.7 → 58.3+0.6
  • Gemini 4 CarbonhighIntelligence59.4 → 60.0+0.6
  • lowIntelligence25.3 → 25.6+0.3
  • Claude Sonnet 5.5mediumIntelligence30.3 → 30.8+0.5
  • Claude Sonnet 5.5highIntelligence41.1 → 41.5+0.4
  • Claude Sonnet 5.5xhighIntelligence49.6 → 49.9+0.3
Show 39 more
  • maxIntelligence54.2 → 54.5+0.3
  • single settingIntelligence46.0 → 46.5+0.5
  • maxIntelligence43.3 → 43.6+0.3
  • highIntelligence29.6 → 29.9+0.3
  • Claude Haiku 5.5xhighIntelligence34.3 → 34.6+0.3
  • Claude Haiku 5.5maxIntelligence36.2 → 36.5+0.3
  • mediumIntelligence30.0 → 30.1+0.1
  • Gemini 3.8 FlashhighIntelligence30.5 → 30.8+0.3
  • lowIntelligence43.0 → 43.3+0.3
  • GPT-6 AstramediumIntelligence51.5 → 51.6+0.1
  • lowIntelligence34.5 → 34.8+0.3
  • GPT-6.1 SolmediumIntelligence46.1 → 46.3+0.2
  • GPT-6.1 SolhighIntelligence50.2 → 50.3+0.1
  • GPT-6.1 SolxhighIntelligence52.3 → 52.5+0.2
  • GPT-6.1 SolmaxIntelligence54.5 → 54.6+0.1
  • lowIntelligence29.2 → 29.4+0.2
  • GPT-6.1 TerramediumIntelligence38.1 → 38.2+0.1
  • GPT-6.1 TerrahighIntelligence42.3 → 42.4+0.1
  • GPT-6.1 TerraxhighIntelligence45.0 → 45.3+0.3
  • GPT-6.1 TerramaxIntelligence51.0 → 51.3+0.3
  • lowIntelligence38.5 → 38.7+0.2
  • Claude Fable 5.1mediumIntelligence43.0 → 43.2+0.2
  • Claude Fable 5.1highIntelligence47.0 → 47.2+0.2
  • Claude Fable 5.1xhighIntelligence49.7 → 50.0+0.3
  • Claude Fable 5.1maxIntelligence49.0 → 49.2+0.2
  • maxIntelligence36.4 → 36.6+0.2
  • lowIntelligence30.9 → 31.1+0.2
  • Grok 4.7highIntelligence35.0 → 35.2+0.2
  • Grok 4.7xhighIntelligence35.3 → 35.5+0.2
  • single settingIntelligence28.9 → 29.1+0.2
  • mediumIntelligence18.3 → 18.5+0.2
  • GPT-6 LunaxhighIntelligence23.3 → 23.5+0.2
  • GPT-6 LunamaxIntelligence27.8 → 27.9+0.1
  • xhighIntelligence24.7 → 24.9+0.2
  • single settingIntelligence18.7 → 18.9+0.2
  • single settingIntelligence33.6 → 33.7+0.1
  • maxIntelligence32.5 → 32.6+0.1
  • single settingIntelligence29.5 → 29.6+0.1
  • highIntelligence26.0 → 26.1+0.1

1 new release · 1 new effort level

New releases 1

  • peak 44.1one setting, $2.47 per task

New effort levels 1

  • lowIntelligence53.7$7.19 per task

1 new release

New releases 1

  • peak 32.2one setting, $3.73 per task

1 score change · 1 price change

Score changes 1

  • xhighIntelligence34.3 → 35.3+1.0

Price changes 1

  • xhighCost per task$7.51 → $7.77+$0.26

1 new release · 1 new effort level

New releases 1

  • peak 5.1one setting, $0.041 per task

New effort levels 1

  • mediumIntelligence57.7$6.59 per task
Load older