Gemini 4 Argon

Composite indexes

• Artificial Analysis Intelligence Index#5 of 151
  53independentAA Intelligence Index v4.3.2, high effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Claude Opus 5.5 at 58

Coding

• Terminal-Bench 4.0#7 of 96
  57.4%vendorGoogle DeepMind: Gemini models ↗
  57%independentTerminal-Bench 4.0 (AA harness), high effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Claude Sonnet 5.5 at 70.6%

• SciCode#4 of 100
  62%independenthigh effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Claude Opus 5.5 at 66.9%

Agents and tool use

• GDPval-AA#12 of 109
  1624independentGDPval-AA, high effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Muse Wonder Full at 1857

• OSWorld-Verified#16 of 21
  69.2%vendorOSWorld 2.0 offlineGoogle DeepMind: Gemini models ↗

  Leader: Kimi K3 at 84.8%

Reasoning and knowledge

• Humanity’s Last Exam#4 of 153
  57%independentno tools (AA), high effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Claude Sonnet 5.5 at 64.5%

• CritPt#10 of 144
  27%independenthigh effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Muse Wonder Full at 33%

Vision and long context

• AA Long Context Reasoning#33 of 144
  80%independenthigh effortArtificial Analysis: Gemini 4 Argon (high) ↗

  Leader: Kimi K3 at 88.7%

Human preference

• LMArena Text#1 of 57
  1525independentgemini-4-argon-high, style control default, preliminary (4,892 votes), snapshot 2026-10-08LMArena Text leaderboard ↗

  Leader: Gemini 4 Argon at 1525

• LMArena WebDev#7 of 58
  1678independentgemini-4-argon-high, snapshot 2026-10-08LMArena WebDev leaderboard ↗

  Leader: Claude Opus 5.5 at 1818

Other published results