priorsBuilder

xAI · released 2026-08-12

Grok 4.6

Measured on 96 benchmarks across 4 reasoning efforts.

General frontier

45.8

#20 of 32 models · best at xhigh

Reasoning efforts in your mix

xhigh45.8#62 of 128
high45.2#67 of 128
medium44.9#71 of 128
low35.2#109 of 128

Your mix’s benchmarks

8 of 8 measured

reasoning

Humanity's Last Exam

587 results
  1. #52xhigh of 58744.1%
  2. #58high of 58742.9%
  3. #70medium of 58742.1%
  4. #169low of 58727.6%

Leader: Opus 5.5 · 61.4%

knowledge

AA-Omniscience

526 results
  1. #31high of 52630.5
  2. #35xhigh of 52629.3
  3. #40medium of 52628.0
  4. #49low of 52625.9

Leader: Opus 5.5 · 46.4

reasoning

ARC-AGI-3

57 results
  1. #26xhigh of 572.1%

Leader: GPT-6 Astra · 99.9%

reasoning

BullshitBench v2

196 results
  1. #35minimal of 19666.0%
  2. #63max of 19656.0%

Leader: Opus 4.8 · 95.0%

coding

DeepSWE v1.1

70 results
  1. #20medium of 7067.5%
  2. #24xhigh of 7066.7%
  3. #28high of 7065.2%
  4. #57low of 7041.6%

Leader: GPT-6 Astra · 74.1%

math

FrontierMath Tier 4 (v2)

70 results
  1. #32xhigh of 7031.7%

Leader: GPT-6.1 Sol · 100.0%

reasoning

SimpleBench

97 results
  1. #14max of 9775.9%

Leader: Opus 5.5 · 88.4%

agentic

Terminal-Bench 4.0

35 results
  1. #28high of 3520.3%

Leader: Opus 5.5 · 64.8%

All 96 benchmarks

show 88 more

agentic

AA-AnalystAgent

44 results
  1. #18high of 4441.3%

Leader: Gemini 3.7 Flash · 60.0%

agentic

AA-Briefcase v1.1

224 results
  1. #25xhigh of 2241545
  2. #27high of 2241535
  3. #39medium of 2241484
  4. #70low of 2241288

Leader: Sonnet 5.5 · 1823

agentic

Agent Arena

50 results
  1. #30xhigh of 501.1%

Leader: Opus 5.5 · 15.5%

agentic

AutomationBench-AA

219 results
  1. #9xhigh of 21967.0%
  2. #10high of 21966.7%
  3. #22medium of 21963.2%
  4. #64low of 21953.4%

Leader: Gemini 4 · 77.5%

agentic

BioMysteryBench

27 results
  1. #10high of 2772.2%

Leader: Sonnet 5.5 · 81.1%

agentic

EnterpriseOps-Gym-AA

50 results
  1. #5high of 5048.3%

Leader: Fable 5 · 51.1%

agentic

Harvey's Legal Agent Benchmark

79 results
  1. #10high of 7915.8%

Leader: Muse Spark 1.2 · 25.4%

agentic

Legal Research Bench

78 results
  1. #10high of 7848.1%

Leader: Muse Spark 1.3 · 55.3%

agentic

SkillsBench

34 results
  1. #12high of 3455.8%

Leader: DeepSeek V4.1 Flash · 69.8%

agentic

Tax Agent Bench

69 results
  1. #14high of 6933.1%

Leader: Fable 5.1 · 49.2%

agentic

Terminal-Bench 4.0 (Vals)

47 results
  1. #28high of 4717.2%

Leader: Opus 5.5 · 65.2%

agentic

Terminal-Bench Science (Vals)

40 results
  1. #13high of 4011.4%

Leader: GPT-6 Astra · 62.9%

agentic

Time Horizon Index: KSP

13 results
  1. #12high of 135.8%

Leader: Opus 5.5 · 91.3%

agentic

Vending-Bench 2

67 results
  1. #10max of 67$9,047

Leader: GPT-6 Astra · $15,515

agentic

τ³-Banking

206 results
  1. #2high of 20650.7%
  2. #18medium of 20644.3%
  3. #20xhigh of 20643.3%
  4. #40low of 20638.1%

Leader: Qwen 3.8 Max · 51.3%

coding

Arena Image-to-WebDev

58 results
  1. #14high of 581601

Leader: Opus 5.5 · 1749

coding

Arena WebDev

122 results
  1. #24high of 1221617

Leader: Opus 5.5 · 1813

coding

Code Migration

77 results
  1. #20high of 7744.6%

Leader: Sonnet 5.5 · 69.8%

coding

FrontierCode 1.1 Extended

130 results
  1. #21high of 13061.3%
  2. #33medium of 13059.6%
  3. #65low of 13055.0%

Leader: Opus 5.5 · 65.3%

coding

FrontierCode 1.1 Main

130 results
  1. #28high of 13048.0%
  2. #41medium of 13046.1%
  3. #70low of 13040.1%

Leader: Opus 5.5 · 54.6%

coding

FrontierSWE v1

17 results
  1. #3max of 1777.9%

Leader: Fable 5 · 88.2%

coding

FrontierSWE v2

21 results
  1. #14max of 2125.3%

Leader: GPT-6 Astra · 65.5%

coding

IOI

44 results
  1. #28high of 4447.6%

Leader: Gemini 4 · 100.0%

coding

LiveCodeBench (Vals)

127 results
  1. #7high of 12788.2%

Leader: Fable 5.1 · 90.5%

coding

ProgramBench v1 Almost Resolved

62 results
  1. #12high of 6219.0%

Leader: Opus 5.5 · 65.0%

coding

ProgramBench v1 Fully Resolved

62 results
  1. #16high of 621.0%

Leader: Opus 5.5 · 18.5%

coding

ProgramBench v1 Raw Pass Rate

62 results
  1. #12high of 6274.0%

Leader: Opus 5.5 · 87.0%

coding

SciCode

226 results
  1. #36high of 22656.5%
  2. #41medium of 22655.9%
  3. #78xhigh of 22653.0%
  4. #113low of 22649.4%

Leader: Opus 5.5 · 66.9%

coding

SWE-bench Verified (Vals)

85 results
  1. #4high of 8595.6%

Leader: Opus 5 · 97.0%

coding

Terminal-Bench 2.1 (Vals)

74 results
  1. #13high of 7478.3%

Leader: Opus 5.5 · 87.6%

coding

Terminal-Bench Science-AA

46 results
  1. #31high of 466.2%

Leader: GPT-6 Astra · 63.3%

coding

Terminal-Bench v2.1-AA

238 results
  1. #11high of 23888.4%
  2. #16xhigh of 23888.0%
  3. #31medium of 23884.3%
  4. #61low of 23875.3%

Leader: Fable 5.1 · 91.4%

coding

Terminal-Bench v4-AA

216 results
  1. #58high of 21621.2%
  2. #64xhigh of 21617.2%
  3. #74medium of 21613.1%
  4. #104low of 2163.0%

Leader: Sonnet 5.5 · 63.6%

coding

Vibe Code Bench 1-100

22 results
  1. #16high of 2214.8%

Leader: Opus 5.5 · 30.4%

coding

Vibe Code Bench v1.1

105 results
  1. #32high of 10576.2%

Leader: Sonnet 5.5 · 92.4%

coding

WeirdML v2

127 results
  1. #34max of 12767.3%

Leader: GPT-6 Astra Pro · 93.6%

composite

AA Intelligence Index v4.3.2

610 results
  1. #39high of 61044.3
  2. #41xhigh of 61044.2
  3. #46medium of 61042.8
  4. #84low of 61035.1

Leader: Opus 5.5 · 57.6

composite

Arena Text

220 results
  1. #62high of 2201454

Leader: Gemini 4 · 1525

composite

Arena Text — Multi-Turn

220 results
  1. #81high of 2201447

Leader: Gemini 4 · 1553

composite

Vals Index

47 results
  1. #23high of 4752.1%

Leader: Gemini 4 · 68.9%

composite

Vals RSI Index v1.1

24 results
  1. #8max of 2425.5%

Leader: Opus 5.5 · 37.3%

composite

Vals RSI Index v1.1: Harness Engineering: Judge

24 results
  1. #3max of 2421.8%

Leader: GPT-6 Sol · 22.8%

composite

Vals RSI Index v1.1: Post-training: Finance Agent

24 results
  1. #12max of 240.0%

Leader: Opus 5.5 · 35.1%

composite

Vals RSI Index v1.1: Pre-training: Compression

24 results
  1. #11max of 2437.8%

Leader: Opus 5.5 · 46.5%

composite

Vals RSI Index v1.1: Pre-training: LM Training

24 results
  1. #7max of 2442.3%

Leader: Opus 5.5 · 53.0%

cost

AA Intelligence Index Cost per Task

177 results
  1. #81low of 177$0.38
  2. #123medium of 177$1.14
  3. #135high of 177$1.48
  4. #142xhigh of 177$1.75

Leader: GPT-6 Luna · $0.00

cost

Cost to Run AA Intelligence Index

177 results
  1. #78low of 177$614
  2. #122medium of 177$1,557
  3. #132high of 177$1,963
  4. #141xhigh of 177$2,242

Leader: GPT-6 Luna · $10.63

creative

MineBench

72 results
  1. #13max of 721850

Leader: GPT-6 Astra Pro · 2281

creative

threejseval

34 results
  1. #18high of 341472
  2. #26medium of 341299
  3. #28low of 341237

Leader: Opus 5.5 · 2079

creative

VoxelBench (text)

55 results
  1. #8xhigh of 552033

Leader: GPT-6 Astra · 2732

economics

CorpFin (Vals)

117 results
  1. #21high of 11766.2%

Leader: Opus 5 · 73.2%

economics

Excel Modeling Benchmark

74 results
  1. #30high of 7462.7%

Leader: Fable 5.1 · 76.7%

economics

Finance Agent v2

78 results
  1. #30high of 7853.7%

Leader: Gemini 4 · 65.4%

economics

GDP.pdf

210 results
  1. #75medium of 21017.8%
  2. #80xhigh of 21017.2%
  3. #83high of 21017.0%
  4. #88low of 21015.8%

Leader: Muse Wonder Full (Max) · 32.6%

economics

GDPval-AA v2.1

288 results
  1. #18xhigh of 2881646
  2. #25high of 2881619
  3. #27medium of 2881613
  4. #71low of 2881409

Leader: Opus 5.5 · 1866

economics

MortgageTax (Vals)

84 results
  1. #44high of 8464.2%

Leader: Opus 5 · 72.1%

economics

TaxEval (Vals)

125 results
  1. #79high of 12571.1%

Leader: Muse Spark 1.2 · 80.4%

knowledge

AA-Omniscience Accuracy

526 results
  1. #70high of 52648.2%
  2. #97low of 52643.3%
  3. #101xhigh of 52643.0%
  4. #106medium of 52641.9%

Leader: Fable 5.1 · 67.2%

knowledge

AA-Omniscience Non-hallucination

526 results
  1. #18medium of 52676.0%
  2. #19xhigh of 52676.0%
  3. #47low of 52669.4%
  4. #59high of 52665.7%

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

knowledge

GPQA Diamond

539 results
  1. #5high of 53994.9%
  2. #15xhigh of 53993.5%
  3. #17medium of 53993.5%
  4. #93low of 53987.9%

Leader: GPT-6 Astra · 96.3%

knowledge

GPQA Diamond (Vals)

122 results
  1. #3high of 12294.7%

Leader: Gemini 3.1 Pro · 95.5%

knowledge

MedCode

94 results
  1. #37high of 9444.7%

Leader: Opus 5 · 63.6%

knowledge

MedScribe

96 results
  1. #21high of 9686.5%

Leader: Opus 5.5 · 91.4%

knowledge

MMLU Pro (Vals)

122 results
  1. #11high of 12289.4%

Leader: Fable 5.1 · 92.4%

knowledge

Public Benefits Bench v1.1

50 results
  1. #17high of 5066.8%

Leader: Opus 5 · 76.9%

long-context

AA-LCR v1.1

523 results
  1. #58xhigh of 52381.0%
  2. #59medium of 52381.0%
  3. #65low of 52380.7%
  4. #70high of 52380.3%

Leader: Kimi K3 · 88.7%

long-context

Arena Document

43 results
  1. #25high of 431452

Leader: Opus 5 · 1516

long-context

MLCR-AA

94 results
  1. #56high of 9412.2%

Leader: Sonnet 5.5 · 75.0%

math

FrontierMath Tiers 1–3 (v2)

115 results
  1. #29xhigh of 11566.0%

Leader: GPT-6.1 Sol · 93.7%

math

ProofBench v1.1

52 results
  1. #31high of 5251.0%

Leader: Sonnet 5.5 · 100.0%

other

Arena Vision

116 results
  1. #44high of 1161262

Leader: Fable 5 · 1308

reasoning

ARC-AGI-1

215 results
  1. #81medium of 21587.5%
  2. #86xhigh of 21587.0%
  3. #87high of 21587.0%
  4. #116low of 21574.8%

Leader: Fable 5 · 98.5%

reasoning

ARC-AGI-2

216 results
  1. #60xhigh of 21667.1%
  2. #66high of 21665.1%
  3. #73medium of 21661.3%
  4. #120low of 21627.6%

Leader: GPT-6 Astra · 95.0%

reasoning

Blueprint-Bench 2

28 results
  1. #12max of 2833.2%

Leader: Gemini 4 · 54.4%

reasoning

CritPt

529 results
  1. #61xhigh of 52919.7%
  2. #72medium of 52917.7%
  3. #78high of 52917.1%
  4. #142low of 5295.7%

Leader: Muse Wonder Full (Max) · 32.9%

reasoning

Furniture Assembly

31 results
  1. #15xhigh of 3140.0%

Leader: Opus 5.5 · 83.3%

reasoning

LegalBench

135 results
  1. #13high of 13586.3%

Leader: Fable 5 · 88.6%

reasoning

MysteryMechanism

27 results
  1. #13high of 2730.6%

Leader: GPT-6 Astra · 53.2%

reasoning

Roboflow Visual Reasoning

102 results
  1. #55high of 10263.8%
  2. #60low of 10261.1%

Leader: GPT-6 Astra · 91.2%

reasoning

SAGE

80 results
  1. #77high of 8028.9%

Leader: Opus 4.7 · 56.1%

reasoning

SimpleQA Verified

47 results
  1. #14max of 4752.9%

Leader: Gemini 3.1 Pro · 77.5%

reliability

Tool-call error rate

171 results
  1. #28max of 1710.3%

Leader: GPT-5.4 Pro · 0.0%

reliability

Uptime (7d)

181 results
  1. #108max of 18199.1%

Leader: GPT-5 Pro · 100.0%

safety

CyberBench v1.1

45 results
  1. #29high of 4567.7%

Leader: Wonder · 80.4%

speed

Latency (Time to First Token)

351 results
  1. #253low of 35111s
  2. #307medium of 35141s
  3. #313high of 35147s
  4. #319xhigh of 35161s

Leader: Gemini 2.5 Flash-Lite · 0.3s

speed

Output Speed (tokens/s)

351 results
  1. #263xhigh of 35159.0
  2. #265high of 35158.7
  3. #271medium of 35157.6
  4. #290low of 35153.8

Leader: Celeris-1 · 1522.5

speed

Provider latency

177 results
  1. #133max of 1772.4s

Leader: Qwen 2.5 7B · 0.2s

speed

Provider throughput

176 results
  1. #117max of 17650.5

Leader: Muse Spark 1.1 · 204.3