Jamba 1.7 Mini

Composite indexes

• Artificial Analysis Intelligence Index#138 of 142
  5independentIntelligence Index v4.3; non-reasoning; estimatedArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: Claude Opus 5.5 at 58

Coding

• LiveCodeBench#56 of 56
  6.1%independentnon-reasoningArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: Fugu at 92.9%

Agents and tool use

• τ²-Bench#79 of 88
  12.6%independenttelecom; non-reasoningArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: GLM-5.2 at 99.1%

Reasoning and knowledge

• GPQA Diamond#139 of 142
  32.2%independentnon-reasoningArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: GPT-6 Astra at 96.1%

• Humanity’s Last Exam#131 of 147
  4.4%independentno tools; non-reasoningArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: Claude Opus 5.5 at 61.4%

• CritPt#132 of 139
  0%independentnon-reasoningArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: GPT-5.6 Sol at 32.3%

Vision and long context

• AA Long Context Reasoning#121 of 139
  16%independentnon-reasoningArtificial Analysis: Jamba 1.7 Mini ↗

  Leader: Kimi K3 at 88.7%

Other published results