Gemma 4 E2B

Composite indexes

• Artificial Analysis Intelligence Index#122 of 151
  8independentv4.3.2, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗
  7independentv4.3.2, non-reasoningArtificial Analysis: Gemma 4 E2B (Non-reasoning) ↗

  Leader: Claude Opus 5.5 at 58

Coding

• Terminal-Bench 2.x#103 of 105
  0.4%independentAA run (Terminal-Bench 2.1), reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: Claude Fable 5.1 at 91.4%

• LiveCodeBench#38 of 57
  44%vendorLiveCodeBench v6Gemma 4 model card ↗

  Leader: Fugu at 92.9%

Agents and tool use

• GDPval-AA#107 of 109
  -203independentGDPval-AA v2.1, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: Muse Wonder Full at 1857

• τ²-Bench#68 of 89
  24.5%vendoraverage over 3 domainsGemma 4 model card ↗
  20.8%independentAA run, telecom, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: GLM-5.2 at 99.1%

Reasoning and knowledge

• GPQA Diamond#133 of 144
  43.4%vendorGemma 4 model card ↗
  43.3%independentAA run, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: GPT-6 Astra at 96.1%

• Humanity’s Last Exam#132 of 153
  4.8%independentAA run, no tools, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: Claude Sonnet 5.5 at 64.5%

• CritPt#102 of 144
  0%independentAA run, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: Muse Wonder Full at 33%

• MMLU-Pro#15 of 15
  60%vendorinstruction-tunedGemma 4 model card ↗

  Leader: Nemotron 3 Ultra 550B A55B at 86.8%

• AIME 2026#15 of 15
  37.5%vendorno toolsGemma 4 model card ↗

  Leader: ERNIE 5.1 at 99.6%

Vision and long context

• AA Long Context Reasoning#125 of 144
  16.3%independentAA-LCR v1.1, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗

  Leader: Kimi K3 at 88.7%

• MMMU-Pro#67 of 72
  44.6%independentAA run, reasoningArtificial Analysis: Gemma 4 E2B (Reasoning) ↗
  44.2%vendorGemma 4 model card ↗

  Leader: Claude Opus 5.5 at 87.7%

Other published results