Benchmarks
More
OverviewBenchmaxxingCompareChartsCost vs CapabilityProviders per ModelProvider explorerGatewaysEU & SovereignJev-class modelsAbout
← All models

Inkling Small

open weights
Thinking Machines · released 2026-07-30 · 4 offers

Output 185 tokens/sFirst token 1.7 sContext 1M tokens

Top 3 cheapest providers (Adjusted $/task)

The same list price can give a different adjusted $/task (caching, token efficiency) — click a price for its inputs.

Within the active global provider, residency and confidentiality filters.

#ProviderAdjusted $/task
1DeepInfra
2BaseTen
3Together

Composite

3 of 7 inputs · 1 from the model family6 radar axes: DesignArena's two boards share one

62.8

AA CodingCoding Agent v1.4AA IntelligenceAA AgenticEpoch ECISoftware ECIDesignArenaAA Coding: percentile 61AA Intelligence: percentile 82AA Agentic: percentile 63Epoch ECI: percentile 55

AA Coding 52.9Coding Agent v1.4 —AA Intelligence 26.1AA Agentic 25.0Epoch ECI 150.2Software ECI —DesignArena —/—

Radar: percentile among all models measured on each input; a gap means not measured.

Benchmark sheet

31 of 140 registered benchmark versions · bars show the percentile among all models measured on each benchmark.

Composite attachments (used in the score, not counted as exact benchmarks):
  • Epoch ECI · attached
Compare this model →

Agentic

  • AA-AnalystAgent5227.5%

    AA-AnalystAgent published 2026-09-10 ↗ · Published board — Tests analyst tasks using agentic Python execution across fourteen domains.

  • AA-Briefcase45912 Elo

    AA-Briefcase published 2026-09-10 ↗ · Published board — Tests multi-week professional knowledge-work projects with linked tasks and large source collections.

  • GDPval-AA v2641,191 Elo

    GDPval-AA v2 v2 ↗ · Published board — Tests professional knowledge-work deliverables across occupations using AA's Stirrup harness.

  • Terminal-Bench v2.1 (AA)5455.1%

    Terminal-Bench v2.1 (AA) v2.1 ↗ · Published board — Tests terminal-based work on the 89-task verified refresh using Terminus 2.

  • Terminal-Bench v4.0 (AA)471.0%

    Terminal-Bench v4.0 (AA) v4.0 ↗ · Published board — Tests terminal-based work on the 66-task release using mini-SWE-agent v2.4.6.

  • Vals Index v2 (Vals AI)731.9%

    Vals Index v2 (Vals AI) v2 ↗ · Published board — GDP-weighted average of agentic model performance across finance, coding and legal tasks.

  • Finance Agent v2 (Vals Index v2)1341.3%

    Finance Agent v2 (Vals Index v2) v2 ↗ · Published board — Multi-step financial reasoning tasks.

  • Terminal-Bench 2.1 (Vals Index v2)1855.1%

    Terminal-Bench 2.1 (Vals Index v2) v2 ↗ · Published board — Command-line interface problem solving, as run by Vals AI.

  • HLAB — Harvey's Legal Agent Benchmark (Vals Index v2)171.7%

    HLAB — Harvey's Legal Agent Benchmark (Vals Index v2) v2 ↗ · Published board — Long-horizon legal work product creation. Vals AI's run of Harvey's Legal Agent Benchmark, reported as accuracy — not comparable with Artificial Analysis' Harvey LAB-AA row.

  • τ²-Bench Airline (OpenRouter run)3160.9%

    τ²-Bench Airline (OpenRouter run) published 2026-09-15 ↗ · Published board — Multi-turn service agents making tool calls under strict policy constraints. (OpenRouter's own reproducible run, as published on openrouter.ai/benchmarks.)

  • AA Agentic Index6325.0

    AA Agentic Index published 2026-09-18 ↗ · Published board — Artificial Analysis publishes one Agentic Index per model, measured on its primary configuration and relayed by OpenRouter's Benchmarks API; the value is attached at family scope on the deterministic representative. Artificial Analysis publishes this board without a version number; we keep the date each result was retained.

Coding

Efficiency

Knowledge

Long-context

  • GDP.pdf (AA)5111.2%

    GDP.pdf (AA) published 2026-09-10 ↗ · Published board — Tests professional reasoning over long PDFs with AA document preparation and grading.

  • AA-LCR v1.17875.7%

    AA-LCR v1.1 v1.1 ↗ · Published board — Tests reasoning across multiple long documents with corrected answer keys and grading.

Reasoning

Science

Tool-use

  • AutomationBench-AA v1.0.6324.8%

    AutomationBench-AA v1.0.6 ↗ · Published board — Tests multi-app SaaS workflows through REST tools on a held-out AutomationBench split.

  • EnterpriseOps-Gym-AA6042.7%

    EnterpriseOps-Gym-AA published 2026-09-10 ↗ · Published board — Tests enterprise workflows through MCP tools against resettable application environments.

  • τ³-Banking (AA) v1.0.15118.8%

    τ³-Banking (AA) v1.0.1 ↗ · Published board — Tests banking support agents that retrieve policies and change account state through tools.

  • Excel Modeling Benchmark (Vals Index v2)631.8%

    Excel Modeling Benchmark (Vals Index v2) v2 ↗ · Published board — Building and editing financial models in spreadsheets.

Vision

Writing

  • The Slop Index8919.4

    The Slop Index published 2026-09-10 ↗ · Published board — Leaderboard scoring how much each model's writing reads like chatbot slop versus pre-AI human writing, blending live arena votes with four mechanical axes.

Unusual results

3 threshold-crossing signals flagged · 31 eligible benchmark families

How flags are calculated

Heuristic screen, not statistical significance: benchmark families are correlated and source uncertainty is unknown. Peer evidence needs ≥ 20 independently measured matched configurations from ≥ 10 distinct model families. A flag needs a directed population z-score of magnitude ≥ 1.5 and a gap of ≥ 1.5 from the leave-one-benchmark-family-out mean z in the same direction, over ≥ 5 other benchmark families.

  • unusually weak Vibe Code Bench (Vals Index v2)
    Why
    Observed: 19.052 percent
    Peer mean 65.32897 · peer sd 21.57746 · n 32 · families 31
    Directed z -2.145 · baseline z -0.023 · gap -2.121 · profile n 30
    19.052 percentmeasuredobserved 2026-09-18vals.ai ↗
    Evidence
    Axis: Vibe Code Bench (Vals Index v2) · v2 · Published board
    Exact value: 19.052 percent
    Observed: 2026-09-18T13:00:43.826192+00:00 · publication date: not recorded
    Observation id: public:162b486be2e2b37ccbe6b5ed
  • unusually weak Excel Modeling Benchmark (Vals Index v2)
    Why
    Observed: 31.762 percent
    Peer mean 59.33688 · peer sd 13.75792 · n 33 · families 32
    Directed z -2.004 · baseline z -0.028 · gap -1.976 · profile n 30
    31.762 percentmeasuredobserved 2026-09-18vals.ai ↗
    Evidence
    Axis: Excel Modeling Benchmark (Vals Index v2) · v2 · Published board
    Exact value: 31.762 percent
    Observed: 2026-09-18T13:00:43.826192+00:00 · publication date: not recorded
    Observation id: public:8e20bb0d9a425b305d70fa57
  • unusually weak Vals Index v2 (Vals AI)
    Why
    Observed: 31.862 percent
    Peer mean 53.34655 · peer sd 11.98829 · n 31 · families 30
    Directed z -1.792 · baseline z -0.035 · gap -1.757 · profile n 30
    31.862 percentmeasured· 0.28 USD per rolloutobserved 2026-09-18vals.ai ↗
    Evidence
    Axis: Vals Index v2 (Vals AI) · v2 · Published board
    Exact value: 31.862 percent
    Published mean cost: 0.281249 USD per rollout
    Observed: 2026-09-18T13:00:43.826192+00:00 · publication date: not recorded
    Observation id: public:e9310c98e8539537e17fda42
Missing coverage · 112 benchmark versions

No result does not mean a zero, or that the model was never tested. Collection failures and disputed versions retain their distinct status.

Token offers by platform · 3 offers (Adjusted $/task)

Click any underlined price to see how it is estimated and where each input comes from. How we calculate adjusted cost.

3 offers within the active global filters; “—” means the catalog is active but no public token price is available.

OpenRouter (3)

DeepInfra
BaseTen
Together
Inkling Small — benchmarks & cost | Benchmark Heaven