priorsBuilder

Thinking Machines · released 2026-07-30

Inkling Small

Measured on 69 benchmarks across 2 reasoning efforts.

General frontier

39.1

#64 of 136 models · best at xhigh

includes estimates: 16% of the mix is measured for this model

Reasoning efforts in your mix

xhigh39.1#152 of 313
default33.7#197 of 313

Your mix’s benchmarks

3 of 8 measured

reasoning

Humanity's Last Exam

587 results
  1. #134default of 58733.3%

Leader: Opus 5.5 · 61.4%

knowledge

AA-Omniscience

526 results
  1. #178default of 526-8.9

Leader: Opus 5.5 · 46.4

math

FrontierMath Tier 4 (v2)

70 results
  1. #52xhigh of 7017.1%

Leader: GPT-6.1 Sol · 100.0%

All 69 benchmarks

show 66 more

agentic

AA-AnalystAgent

44 results
  1. #25default of 4427.5%

Leader: Gemini 3.7 Flash · 60.0%

agentic

AA-Briefcase v1.1

224 results
  1. #134default of 224912

Leader: Sonnet 5.5 · 1823

agentic

Agent Arena

50 results
  1. #48xhigh of 50-10.3%

Leader: Opus 5.5 · 15.5%

agentic

AutomationBench-AA

219 results
  1. #162default of 2194.8%

Leader: Gemini 4 · 77.5%

agentic

EnterpriseOps-Gym-AA

50 results
  1. #22default of 5042.7%

Leader: Fable 5 · 51.1%

agentic

Harvey's Legal Agent Benchmark

79 results
  1. #52xhigh of 791.7%

Leader: Muse Spark 1.2 · 25.4%

agentic

Legal Research Bench

78 results
  1. #51xhigh of 7825.5%

Leader: Muse Spark 1.3 · 55.3%

agentic

SkillsBench

34 results
  1. #31xhigh of 3433.6%

Leader: DeepSeek V4.1 Flash · 69.8%

agentic

Tax Agent Bench

69 results
  1. #49xhigh of 6915.4%

Leader: Fable 5.1 · 49.2%

agentic

Terminal-Bench 4.0 (Vals)

47 results
  1. #46xhigh of 470.0%

Leader: Opus 5.5 · 65.2%

agentic

τ³-Banking

206 results
  1. #104default of 20618.8%

Leader: Qwen 3.8 Max · 51.3%

coding

Arena Image-to-WebDev

58 results
  1. #54xhigh of 581410

Leader: Opus 5.5 · 1749

coding

Arena WebDev

122 results
  1. #79xhigh of 1221409

Leader: Opus 5.5 · 1813

coding

Code Migration

77 results
  1. #58xhigh of 7713.7%

Leader: Sonnet 5.5 · 69.8%

coding

IOI

44 results
  1. #43xhigh of 449.3%

Leader: Gemini 4 · 100.0%

coding

LiveCodeBench (Vals)

127 results
  1. #26xhigh of 12785.9%

Leader: Fable 5.1 · 90.5%

coding

ProgramBench v1 Almost Resolved

62 results
  1. #51xhigh of 620.5%

Leader: Opus 5.5 · 65.0%

coding

ProgramBench v1 Fully Resolved

62 results
  1. #26xhigh of 620.5%

Leader: Opus 5.5 · 18.5%

coding

ProgramBench v1 Raw Pass Rate

62 results
  1. #46xhigh of 6242.5%

Leader: Opus 5.5 · 87.0%

coding

SciCode

226 results
  1. #111default of 22649.7%

Leader: Opus 5.5 · 66.9%

coding

SWE-bench Verified (Vals)

85 results
  1. #19xhigh of 8582.2%

Leader: Opus 5 · 97.0%

coding

Terminal-Bench 2.1 (Vals)

74 results
  1. #47xhigh of 7455.1%

Leader: Opus 5.5 · 87.6%

coding

Terminal-Bench Science-AA

46 results
  1. #42default of 460.0%

Leader: GPT-6 Astra · 63.3%

coding

Terminal-Bench v2.1-AA

238 results
  1. #109default of 23855.1%

Leader: Fable 5.1 · 91.4%

coding

Terminal-Bench v4-AA

216 results
  1. #128default of 2161.0%

Leader: Sonnet 5.5 · 63.6%

coding

Vibe Code Bench v1.1

105 results
  1. #78xhigh of 10519.1%

Leader: Sonnet 5.5 · 92.4%

composite

AA Intelligence Index v4.3.2

610 results
  1. #154default of 61025.7

Leader: Opus 5.5 · 57.6

composite

Arena Text

220 results
  1. #119xhigh of 2201405

Leader: Gemini 4 · 1525

composite

Arena Text — Multi-Turn

220 results
  1. #109xhigh of 2201416

Leader: Gemini 4 · 1553

composite

Vals Index

47 results
  1. #46xhigh of 4725.5%

Leader: Gemini 4 · 68.9%

composite

Vals Multimodal Index v1.2

33 results
  1. #26xhigh of 3350.1%

Leader: Fable 5 · 74.2%

cost

AA Intelligence Index Cost per Task

177 results
  1. #36default of 177$0.09

Leader: GPT-6 Luna · $0.00

cost

Cost to Run AA Intelligence Index

177 results
  1. #45default of 177$241

Leader: GPT-6 Luna · $10.63

economics

CorpFin (Vals)

117 results
  1. #6xhigh of 11769.6%

Leader: Opus 5 · 73.2%

economics

Excel Modeling Benchmark

74 results
  1. #62xhigh of 7431.8%

Leader: Fable 5.1 · 76.7%

economics

Finance Agent v2

78 results
  1. #60xhigh of 7841.3%

Leader: Gemini 4 · 65.4%

economics

GDP.pdf

210 results
  1. #119default of 21011.2%

Leader: Muse Wonder Full (Max) · 32.6%

economics

GDPval-AA v2.1

288 results
  1. #128default of 2881123

Leader: Opus 5.5 · 1866

economics

MortgageTax (Vals)

84 results
  1. #52xhigh of 8462.3%

Leader: Opus 5 · 72.1%

economics

TaxEval (Vals)

125 results
  1. #18xhigh of 12575.5%

Leader: Muse Spark 1.2 · 80.4%

knowledge

AA-Omniscience Accuracy

526 results
  1. #162default of 52633.2%

Leader: Fable 5.1 · 67.2%

knowledge

AA-Omniscience Non-hallucination

526 results
  1. #174default of 52637.0%

Leader: MiniCPM5-1B (Non-reasoning) · 99.1%

knowledge

GPQA Diamond

539 results
  1. #77default of 53989.5%

Leader: GPT-6 Astra · 96.3%

knowledge

GPQA Diamond (Vals)

122 results
  1. #57xhigh of 12283.6%

Leader: Gemini 3.1 Pro · 95.5%

knowledge

Harvey LAB-AA v1.1

27 results
  1. #25default of 270.0%

Leader: Muse Wonder Full (Max) · 13.1%

knowledge

MedCode

94 results
  1. #67xhigh of 9437.9%

Leader: Opus 5 · 63.6%

knowledge

MedScribe

96 results
  1. #36xhigh of 9684.1%

Leader: Opus 5.5 · 91.4%

knowledge

MMLU Pro (Vals)

122 results
  1. #56xhigh of 12285.6%

Leader: Fable 5.1 · 92.4%

knowledge

MMMU-Pro

265 results
  1. #112default of 26574.0%

Leader: Opus 5.5 · 87.7%

knowledge

Public Benefits Bench v1.1

50 results
  1. #34xhigh of 5059.4%

Leader: Opus 5 · 76.9%

long-context

AA-LCR v1.1

523 results
  1. #152default of 52375.7%

Leader: Kimi K3 · 88.7%

math

FrontierMath Tiers 1–3 (v2)

115 results
  1. #51xhigh of 11546.3%

Leader: GPT-6.1 Sol · 93.7%

math

ProofBench v1.1

52 results
  1. #49xhigh of 526.0%

Leader: Sonnet 5.5 · 100.0%

other

Arena Vision

116 results
  1. #75xhigh of 1161208

Leader: Fable 5 · 1308

reasoning

ARC-AGI-1

215 results
  1. #100xhigh of 21584.0%
  2. #109high of 21578.0%
  3. #128medium of 21567.0%
  4. #151low of 21552.5%
  5. #155minimal of 21548.5%
  6. #205none of 2155.5%

Leader: Fable 5 · 98.5%

reasoning

ARC-AGI-2

216 results
  1. #104xhigh of 21640.1%
  2. #113high of 21633.1%
  3. #131medium of 21613.6%
  4. #154low of 2164.9%
  5. #164minimal of 2164.0%
  6. #215none of 2160.0%

Leader: GPT-6 Astra · 95.0%

reasoning

CritPt

529 results
  1. #130default of 5298.3%

Leader: Muse Wonder Full (Max) · 32.9%

reasoning

LegalBench

135 results
  1. #61xhigh of 13583.0%

Leader: Fable 5 · 88.6%

reasoning

SAGE

80 results
  1. #69xhigh of 8032.7%

Leader: Opus 4.7 · 56.1%

reliability

Tool-call error rate

171 results
  1. #149xhigh of 1715.5%

Leader: GPT-5.4 Pro · 0.0%

reliability

Uptime (7d)

181 results
  1. #120xhigh of 18198.8%

Leader: GPT-5 Pro · 100.0%

safety

CyberBench v1.1

45 results
  1. #34xhigh of 4564.1%

Leader: Wonder · 80.4%

speed

Latency (Time to First Token)

351 results
  1. #169default of 3512.2s

Leader: Gemini 2.5 Flash-Lite · 0.3s

speed

Output Speed (tokens/s)

351 results
  1. #34default of 351207.4

Leader: Celeris-1 · 1522.5

speed

Provider latency

177 results
  1. #24xhigh of 1770.6s

Leader: Qwen 2.5 7B · 0.2s

speed

Provider throughput

176 results
  1. #18xhigh of 176114.0

Leader: Muse Spark 1.1 · 204.3