priorsBuilder

Anthropic · released 2025-10-01

Haiku 4.5

Measured on 79 benchmarks.

General frontier

Not ranked in this mix — too little measured data for its benchmarks.

Your mix’s benchmarks

3 of 8 measured

reasoning

Humanity's Last Exam

587 results
  1. #308default of 58710.4%
  2. #505none of 5874.2%

Leader: Opus 5.5 · 61.4%

knowledge

AA-Omniscience

526 results
  1. #154default of 526-4.4
  2. #171none of 526-7.6

Leader: Opus 5.5 · 46.4

reasoning

BullshitBench v2

196 results
  1. #18high of 19677.0%
  2. #27none of 19671.0%

Leader: Opus 4.8 · 95.0%

All 79 benchmarks

show 76 more

agentic

AA-AnalystAgent

44 results
  1. #36default of 4415.0%

Leader: Gemini 3.7 Flash · 60.0%

agentic

AA-Briefcase v1.1

224 results
  1. #168default of 224616

Leader: Sonnet 5.5 · 1823

agentic

AutomationBench (Zapier)

115 results
  1. #111high of 1150.5%

Leader: Gemini 4 · 51.3%

agentic

AutomationBench-AA

219 results
  1. #167default of 2193.2%

Leader: Gemini 4 · 77.5%

agentic

EnterpriseOps-Gym-AA

50 results
  1. #42default of 5029.4%

Leader: Fable 5 · 51.1%

agentic

Harvey's Legal Agent Benchmark

79 results
  1. #58high of 790.8%

Leader: Muse Spark 1.2 · 25.4%

agentic

ITBench-AA

45 results
  1. #35default of 4527.3%

Leader: GPT-5.6 Sol · 56.2%

agentic

Legal Research Bench

78 results
  1. #68high of 7810.6%

Leader: Muse Spark 1.3 · 55.3%

agentic

Tax Agent Bench

69 results
  1. #64high of 695.9%

Leader: Fable 5.1 · 49.2%

agentic

Terminal-Bench 2.0

40 results
  1. #27high of 4035.5%

Leader: GPT-5.5 · 84.7%

agentic

Vending-Bench 2

67 results
  1. #58high of 67$459

Leader: GPT-6 Astra · $15,515

agentic

τ²-Bench Telecom

393 results
  1. #190default of 39354.7%
  2. #246none of 39332.5%

Leader: GLM 5.2 · 99.1%

agentic

τ³-Banking

206 results
  1. #153default of 2069.3%

Leader: Qwen 3.8 Max · 51.3%

coding

Arena WebDev

122 results
  1. #105high of 1221330

Leader: Opus 5.5 · 1813

coding

Code Migration

77 results
  1. #65high of 777.6%

Leader: Sonnet 5.5 · 69.8%

coding

DeepSWE v1

29 results
  1. #28high of 290.2%

Leader: GPT-5.5 · 70.0%

coding

LiveCodeBench

282 results
  1. #100default of 28261.5%
  2. #127none of 28251.1%

Leader: Gemini 3 Pro · 91.7%

coding

LiveCodeBench (Vals)

127 results
  1. #116high of 12741.2%

Leader: Fable 5.1 · 90.5%

coding

ProgramBench v1 Almost Resolved

62 results
  1. #52high of 620.0%

Leader: Opus 5.5 · 65.0%

coding

ProgramBench v1 Fully Resolved

62 results
  1. #33high of 620.0%

Leader: Opus 5.5 · 18.5%

coding

ProgramBench v1 Raw Pass Rate

62 results
  1. #53high of 6225.4%

Leader: Opus 5.5 · 87.0%

coding

SciCode

226 results
  1. #157default of 22642.2%

Leader: Opus 5.5 · 66.9%

coding

SWE-bench Verified (Vals)

85 results
  1. #68high of 8566.6%

Leader: Opus 5 · 97.0%

coding

Terminal-Bench 2.0 (Vals)

63 results
  1. #35high of 6338.2%

Leader: GPT-5.5 · 73.2%

coding

Terminal-Bench 2.1 (Vals)

74 results
  1. #64high of 7443.8%

Leader: Opus 5.5 · 87.6%

coding

Terminal-Bench Hard-AA

387 results
  1. #126default of 38727.3%
  2. #129none of 38727.3%

Leader: GPT-5.6 Sol · 65.9%

coding

Terminal-Bench v2.1-AA

238 results
  1. #133default of 23844.2%

Leader: Fable 5.1 · 91.4%

coding

Terminal-Bench v4-AA

216 results
  1. #168default of 2160.0%

Leader: Sonnet 5.5 · 63.6%

coding

Vibe Code Bench v1.1

105 results
  1. #87high of 10511.4%

Leader: Sonnet 5.5 · 92.4%

composite

AA Intelligence Index v4.3.2

610 results
  1. #262default of 61016.9
  2. #279none of 61015.4

Leader: Opus 5.5 · 57.6

composite

Arena Text

220 results
  1. #110high of 2201414

Leader: Gemini 4 · 1525

composite

Arena Text — Multi-Turn

220 results
  1. #100high of 2201425

Leader: Gemini 4 · 1553

composite

Vals Multimodal Index v1.2

33 results
  1. #30high of 3342.9%

Leader: Fable 5 · 74.2%

cost

AA Intelligence Index Cost per Task

177 results
  1. #67default of 177$0.28

Leader: GPT-6 Luna · $0.00

cost

Cost to Run AA Intelligence Index

177 results
  1. #76default of 177$594

Leader: GPT-6 Luna · $10.63

economics

CorpFin (Vals)

117 results
  1. #66high of 11760.6%

Leader: Opus 5 · 73.2%

economics

Excel Modeling Benchmark

74 results
  1. #68high of 7422.7%

Leader: Fable 5.1 · 76.7%

economics

Finance Agent v2

78 results
  1. #69high of 7831.0%

Leader: Gemini 4 · 65.4%

economics

GDP.pdf

210 results
  1. #169default of 2103.8%

Leader: Muse Wonder Full (Max) · 32.6%

economics

GDPval-AA v2.1

288 results
  1. #193default of 288732

Leader: Opus 5.5 · 1866

economics

MortgageTax (Vals)

84 results
  1. #53high of 8462.2%

Leader: Opus 5 · 72.1%

economics

Public Benefits Bench v1 (Vals)

13 results
  1. #12high of 1349.5%

Leader: Fable 5 · 71.7%

economics

TaxEval (Vals)

125 results
  1. #95high of 12567.5%

Leader: Muse Spark 1.2 · 80.4%

instruction

IFBench

399 results
  1. #152default of 39954.3%
  2. #237none of 39942.0%

Leader: Grok 4.3 · 83.3%

knowledge

AA-Omniscience Accuracy

526 results
  1. #331default of 52618.0%
  2. #411none of 52614.4%

Leader: Fable 5.1 · 67.2%

knowledge

AA-Omniscience Non-hallucination

526 results
  1. #25none of 52674.3%
  2. #32default of 52672.7%

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

knowledge

CaseLaw (Vals)

53 results
  1. #30high of 5356.5%

Leader: Grok 4.3 · 79.3%

knowledge

Global-MMLU-Lite

118 results
  1. #48default of 11883.4%

Leader: Gemini 3.1 Pro · 93.2%

knowledge

GPQA Diamond

539 results
  1. #312default of 53967.2%
  2. #335none of 53964.6%

Leader: GPT-6 Astra · 96.3%

knowledge

GPQA Diamond (Vals)

122 results
  1. #86high of 12272.2%

Leader: Gemini 3.1 Pro · 95.5%

knowledge

MedCode

94 results
  1. #79high of 9432.7%

Leader: Opus 5 · 63.6%

knowledge

MedQA (Vals)

80 results
  1. #62high of 8079.6%

Leader: o1 · 96.5%

knowledge

MedScribe

96 results
  1. #30high of 9685.2%

Leader: Opus 5.5 · 91.4%

knowledge

MMLU Pro (Vals)

122 results
  1. #93high of 12278.7%

Leader: Fable 5.1 · 92.4%

knowledge

MMLU-Pro

285 results
  1. #96none of 28580.0%
  2. #139default of 28576.0%

Leader: Gemini 3 Pro · 89.8%

knowledge

MMMU Pro (Vals)

81 results
  1. #80high of 8146.1%

Leader: Fable 5.1 · 90.6%

knowledge

MMMU-Pro

265 results
  1. #205default of 26558.6%
  2. #214none of 26555.1%

Leader: Opus 5.5 · 87.7%

knowledge

Public Benefits Bench v1.1

50 results
  1. #42high of 5054.3%

Leader: Opus 5 · 76.9%

long-context

AA-LCR v1.1

523 results
  1. #164default of 52374.3%
  2. #308none of 52349.7%

Leader: Kimi K3 · 88.7%

long-context

Arena Document

43 results
  1. #36high of 431420

Leader: Opus 5 · 1516

long-context

MLCR-AA

94 results
  1. #76default of 946.1%
  2. #79none of 944.4%

Leader: Sonnet 5.5 · 75.0%

math

AIME (Vals)

82 results
  1. #44high of 8282.7%

Leader: Gemini 3.1 Pro · 98.1%

math

AIME 2025

228 results
  1. #47default of 22883.7%
  2. #136none of 22839.0%

Leader: GPT-5.2 · 99.0%

math

MGSM (Vals)

62 results
  1. #18high of 6292.1%

Leader: Opus 4.5 · 95.2%

other

T3 Code usage share

85 results
  1. #18high of 850.1%

Leader: Opus 5.5 · 29.1%

reasoning

ARC-AGI-1

215 results
  1. #156high of 21547.7%

Leader: Fable 5 · 98.5%

reasoning

ARC-AGI-2

216 results
  1. #162high of 2164.0%

Leader: GPT-6 Astra · 95.0%

reasoning

CritPt

529 results
  1. #323default of 5290.0%
  2. #328none of 5290.0%

Leader: Muse Wonder Full (Max) · 32.9%

reasoning

LegalBench

135 results
  1. #78high of 13581.2%

Leader: Fable 5 · 88.6%

reasoning

SAGE

80 results
  1. #70high of 8031.8%

Leader: Opus 4.7 · 56.1%

reliability

Tool-call error rate

171 results
  1. #96high of 1711.5%

Leader: GPT-5.4 Pro · 0.0%

reliability

Uptime (7d)

181 results
  1. #60high of 18199.9%

Leader: GPT-5 Pro · 100.0%

speed

Latency (Time to First Token)

351 results
  1. #5none of 3510.5s
  2. #284default of 35122s

Leader: Gemini 2.5 Flash-Lite · 0.3s

speed

Output Speed (tokens/s)

351 results
  1. #175default of 35194.7
  2. #190none of 35187.3

Leader: Celeris-1 · 1522.5

speed

Provider latency

177 results
  1. #27high of 1770.6s

Leader: Qwen 2.5 7B · 0.2s

speed

Provider throughput

176 results
  1. #55high of 17681.0

Leader: Muse Spark 1.1 · 204.3