priorsBuilder

DeepSeek · released 2026-08-13

DeepSeek V4 Pro 0813

Measured on 64 benchmarks across 4 reasoning efforts.

General frontier

42.5

#50 of 136 models · best at max

includes estimates: 24% of the mix is measured for this model

Reasoning efforts in your mix

max42.5#125 of 313
low35.1#186 of 313
max33.9#193 of 313
none20.1#274 of 313

Your mix’s benchmarks

4 of 8 measured

reasoning

Humanity's Last Exam

587 results
  1. #76max of 58741.0%
  2. #304none of 58710.6%

Leader: Opus 5.5 · 61.4%

knowledge

AA-Omniscience

526 results
  1. #125max of 5260.8
  2. #205none of 526-12.4

Leader: Opus 5.5 · 46.4

reasoning

BullshitBench v2

196 results
  1. #109max of 19635.0%
  2. #117low of 19632.0%

Leader: Opus 4.8 · 95.0%

math

FrontierMath Tier 4 (v2)

70 results
  1. #39max of 7026.8%

Leader: GPT-6.1 Sol · 100.0%

All 64 benchmarks

show 60 more

agentic

AA-Briefcase v1.1

224 results
  1. #80max of 2241257
  2. #129none of 224928

Leader: Sonnet 5.5 · 1823

agentic

AutomationBench-AA

219 results
  1. #49max of 21956.7%
  2. #105none of 21934.9%

Leader: Gemini 4 · 77.5%

agentic

EnterpriseOps-Gym-AA

50 results
  1. #4max of 5049.6%

Leader: Fable 5 · 51.1%

agentic

Harvey's Legal Agent Benchmark

79 results
  1. #27max of 797.5%

Leader: Muse Spark 1.2 · 25.4%

agentic

Legal Research Bench

78 results
  1. #28max of 7840.9%

Leader: Muse Spark 1.3 · 55.3%

agentic

SkillsBench

34 results
  1. #15max of 3453.8%

Leader: DeepSeek V4.1 Flash · 69.8%

agentic

Tax Agent Bench

69 results
  1. #32max of 6926.8%

Leader: Fable 5.1 · 49.2%

agentic

Terminal-Bench 4.0 (Vals)

47 results
  1. #30max of 4714.1%

Leader: Opus 5.5 · 65.2%

agentic

Terminal-Bench Science (Vals)

40 results
  1. #29max of 404.3%

Leader: GPT-6 Astra · 62.9%

agentic

τ³-Banking

206 results
  1. #32max of 20639.6%

Leader: Qwen 3.8 Max · 51.3%

coding

Code Migration

77 results
  1. #28max of 7741.5%

Leader: Sonnet 5.5 · 69.8%

coding

FrontierCode 1.1 Extended

130 results
  1. #102high of 13044.5%

Leader: Opus 5.5 · 65.3%

coding

FrontierCode 1.1 Main

130 results
  1. #103high of 13028.5%

Leader: Opus 5.5 · 54.6%

coding

IOI

44 results
  1. #25max of 4451.6%

Leader: Gemini 4 · 100.0%

coding

LiveCodeBench (Vals)

127 results
  1. #13max of 12787.5%

Leader: Fable 5.1 · 90.5%

coding

ProgramBench v1 Almost Resolved

62 results
  1. #23max of 6213.5%

Leader: Opus 5.5 · 65.0%

coding

ProgramBench v1 Fully Resolved

62 results
  1. #38max of 620.0%

Leader: Opus 5.5 · 18.5%

coding

ProgramBench v1 Raw Pass Rate

62 results
  1. #21max of 6270.1%

Leader: Opus 5.5 · 87.0%

coding

SciCode

226 results
  1. #96max of 22651.0%
  2. #172none of 22639.9%

Leader: Opus 5.5 · 66.9%

coding

SWE-bench Verified (Vals)

85 results
  1. #2max of 8596.4%

Leader: Opus 5 · 97.0%

coding

Terminal-Bench 2.1 (Vals)

74 results
  1. #50max of 7454.7%

Leader: Opus 5.5 · 87.6%

coding

Terminal-Bench Science-AA

46 results
  1. #33max of 465.7%

Leader: GPT-6 Astra · 63.3%

coding

Terminal-Bench v2.1-AA

238 results
  1. #49max of 23878.7%

Leader: Fable 5.1 · 91.4%

coding

Terminal-Bench v4-AA

216 results
  1. #72max of 21614.1%
  2. #121none of 2161.5%

Leader: Sonnet 5.5 · 63.6%

coding

Vibe Code Bench 1-100

22 results
  1. #12max of 2217.5%

Leader: Opus 5.5 · 30.4%

coding

Vibe Code Bench v1.1

105 results
  1. #21max of 10582.3%

Leader: Sonnet 5.5 · 92.4%

coding

WeirdML v2

127 results
  1. #35max of 12766.2%

Leader: GPT-6 Astra Pro · 93.6%

composite

AA Intelligence Index v4.3.2

610 results
  1. #82max of 61036.0
  2. #225none of 61020.4

Leader: Opus 5.5 · 57.6

composite

Vals Index

47 results
  1. #37max of 4747.6%

Leader: Gemini 4 · 68.9%

cost

AA Intelligence Index Cost per Task

177 results
  1. #77none of 177$0.36
  2. #99max of 177$0.67

Leader: GPT-6 Luna · $0.00

cost

Cost to Run AA Intelligence Index

177 results
  1. #66none of 177$451
  2. #102max of 177$1,122

Leader: GPT-6 Luna · $10.63

economics

CorpFin (Vals)

117 results
  1. #29max of 11765.4%

Leader: Opus 5 · 73.2%

economics

Excel Modeling Benchmark

74 results
  1. #51max of 7452.8%

Leader: Fable 5.1 · 76.7%

economics

Finance Agent v2

78 results
  1. #42max of 7850.4%

Leader: Gemini 4 · 65.4%

economics

GDP.pdf

210 results
  1. #114max of 21011.4%
  2. #171none of 2103.6%

Leader: Muse Wonder Full (Max) · 32.6%

economics

GDPval-AA v2.1

288 results
  1. #60max of 2881449
  2. #147none of 2881058

Leader: Opus 5.5 · 1866

economics

TaxEval (Vals)

125 results
  1. #53max of 12573.1%

Leader: Muse Spark 1.2 · 80.4%

knowledge

AA-Omniscience Accuracy

526 results
  1. #65max of 52649.1%
  2. #111none of 52641.4%

Leader: Fable 5.1 · 67.2%

knowledge

AA-Omniscience Non-hallucination

526 results
  1. #447none of 5268.1%
  2. #496max of 5265.2%

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

knowledge

GPQA Diamond

539 results
  1. #30max of 53992.8%

Leader: GPT-6 Astra · 96.3%

knowledge

GPQA Diamond (Vals)

122 results
  1. #17max of 12292.4%

Leader: Gemini 3.1 Pro · 95.5%

knowledge

MedCode

94 results
  1. #47max of 9442.5%

Leader: Opus 5 · 63.6%

knowledge

MedScribe

96 results
  1. #54max of 9680.2%

Leader: Opus 5.5 · 91.4%

knowledge

MMLU Pro (Vals)

122 results
  1. #35max of 12287.0%

Leader: Fable 5.1 · 92.4%

long-context

AA-LCR v1.1

523 results
  1. #73max of 52380.3%
  2. #304none of 52350.7%

Leader: Kimi K3 · 88.7%

long-context

MLCR-AA

94 results
  1. #38max of 9417.8%

Leader: Sonnet 5.5 · 75.0%

math

FrontierMath Tiers 1–3 (v2)

115 results
  1. #33max of 11564.6%

Leader: GPT-6.1 Sol · 93.7%

math

ProofBench v1.1

52 results
  1. #32max of 5250.0%

Leader: Sonnet 5.5 · 100.0%

reasoning

ARC-AGI-1

215 results
  1. #67low of 21590.5%
  2. #71max of 21590.0%
  3. #84high of 21587.2%
  4. #193none of 21513.0%

Leader: Fable 5 · 98.5%

reasoning

ARC-AGI-2

216 results
  1. #74max of 21661.3%
  2. #79high of 21659.7%
  3. #87low of 21656.3%
  4. #194none of 2160.8%

Leader: GPT-6 Astra · 95.0%

reasoning

CritPt

529 results
  1. #70max of 52918.0%
  2. #318none of 5290.0%

Leader: Muse Wonder Full (Max) · 32.9%

reasoning

LegalBench

135 results
  1. #69max of 13582.4%

Leader: Fable 5 · 88.6%

reliability

Tool-call error rate

171 results
  1. #69max of 1710.9%

Leader: GPT-5.4 Pro · 0.0%

reliability

Uptime (7d)

181 results
  1. #107max of 18199.2%

Leader: GPT-5 Pro · 100.0%

safety

CyberBench v1.1

45 results
  1. #14max of 4572.9%

Leader: Wonder · 80.4%

safety

SRE Bench

32 results
  1. #23max of 321.9%

Leader: GPT-6 Astra · 56.9%

speed

Latency (Time to First Token)

351 results
  1. #122none of 3511.6s
  2. #126max of 3511.6s

Leader: Gemini 2.5 Flash-Lite · 0.3s

speed

Output Speed (tokens/s)

351 results
  1. #174max of 35194.8
  2. #177none of 35194.0

Leader: Celeris-1 · 1522.5

speed

Provider latency

177 results
  1. #90max of 1771.3s

Leader: Qwen 2.5 7B · 0.2s

speed

Provider throughput

176 results
  1. #100max of 17656.0

Leader: Muse Spark 1.1 · 204.3