Inkling

Composite indexes

• Artificial Analysis Intelligence Index#55 of 142
  25independentIntelligence Index v4.3; xhigh effortArtificial Analysis: Inkling (xhigh) ↗

  Leader: Claude Opus 5.5 at 58

Coding

• SciCode#51 of 94
  47%independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗
  46.1%vendorThinking Machines Lab: Introducing Inkling-Small ↗

  Leader: Claude Opus 5.5 at 66.9%

• Terminal-Bench 4.0#47 of 88
  1%independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗

  Leader: Claude Mythos 5.1 at 60.9%

• Terminal-Bench 2.x#55 of 103
  63.8%vendorTerminal-Bench 2.1, best/internal harness; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗
  55.1%independentTerminal-Bench 2.1; xhigh effortArtificial Analysis: Inkling (xhigh) ↗

  Leader: Claude Fable 5.1 at 91.4%

• SWE-bench Verified#6 of 16
  77.6%vendorbash-only harness; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: Claude Sonnet 5 at 85.2%

• SWE-bench Pro#20 of 31
  54.3%vendorpublic set; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: Claude Opus 5.5 at 89.9%

Agents and tool use

• GDPval-AA#47 of 103
  1238vendorGDPval-AA v2; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗
  1064independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗

  Leader: Claude Opus 5.5 at 1846

• MCP Atlas#5 of 9
  76%vendoreffort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: Kimi K3 at 84.2%

• BrowseComp#8 of 14
  77.1%vendorwith context management; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: GPT-6 Astra at 91.5%

Reasoning and knowledge

• GPQA Diamond#50 of 142
  87.2%independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗
  87.2%vendoreffort 0.99 (AA-reported value cited by vendor)Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: GPT-6 Astra at 96.1%

• Humanity’s Last Exam#55 of 147
  46%vendorwith tools; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗
  31.9%independentno tools; xhigh effortArtificial Analysis: Inkling (xhigh) ↗
  29.7%vendortext only, no tools; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: Claude Opus 5.5 at 61.4%

• CritPt#49 of 139
  5.4%independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗
  5.4%vendorThinking Machines Lab: Introducing Inkling-Small ↗

  Leader: GPT-5.6 Sol at 32.3%

• AIME 2026#3 of 13
  97.1%vendoreffort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: ERNIE 5.1 at 99.6%

• HMMT February 2026#4 of 8
  86.3%vendorHMMT Feb 2026Thinking Machines Lab: Introducing Inkling-Small ↗

  Leader: GLM-5.2 at 92.5%

• ARC-AGI-2#19 of 25
  36.5%independentsemi-private; ARC Prize verified; $0.64/taskARC Prize leaderboard (evaluations data) ↗

  Leader: GPT-6 Astra at 95%

Vision and long context

• AA Long Context Reasoning#51 of 139
  77.3%independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗

  Leader: Kimi K3 at 88.7%

• MMMU-Pro#41 of 72
  73.5%independentxhigh effortArtificial Analysis: Inkling (xhigh) ↗
  73.5%vendorstandard 10 options; effort 0.99Thinking Machines Lab: Inkling: Our Open-Weights Model ↗

  Leader: Claude Opus 5.5 at 87.7%

Other published results