Models Agents EvalsIndustry

AgMoDB by @mistakeknot

DeepSeek V3.2 Exp (Reasoning) | AgMoDB

DeepSeek V3.2 Exp (Reasoning)

September 29, 2025Last synced Apr 7, 2026, 4:00 PM

Blended Price

$0.32/M

Input Price

$0.28/M

Output Price

$0.42/M

Speed

48 tok/s

TTFT

1.24s

Benchmark Scores

Intelligence Index

External Benchmarks

Aider Polyglot

How DeepSeek V3.2 Exp (Reasoning) Compares

Axes

X Axis

Y Axis

Bubble Size

Filters

Blended Price (USD)$0.00 – $30.0

AgMoBench Overall≥ 3.5

Providers

ai21-labsalibabaanthropicawsazurebaiducoheredeepseek

Show quadrants

anthropic

openai

google

meta

mistral

nvidia

kimi

xai

azure

deepseek

aws

cohere

baidu

ai21-labs

zai

reka-ai

xiaomi

alibaba

minimax

ibm

Bubble size = Context Window

16384.00

2000000.00

Compare with other models

32.9

Coding Index

33.3

Math Index

87.7

MMLU Pro

0.8 / 100

GPQA Diamond

0.8 / 100

HLE

0.1 / 100

LiveCodeBench

0.8 / 100

SciCode

0.4 / 100

MATH-500

— / 100

AIME

— / 30

AIME 2025

0.9 / 30

IFBench

0.5 / 100

LCR

0.7 / 100

Terminal-Bench Hard

0.3 / 100

τ²-Bench

0.3 / 100

aider

70.2 / 100

AA-Omniscience AccuracyPredicted

38.5 / 100

AA-Omniscience Hallucination RatePredicted

95.2 / 100

AIMEPredicted

0.8 / 30

AlpacaEval 2.0 LCPredicted

34.7 / 100

AlpacaEval 2.0 RawPredicted

27.0 / 100

ARC-AGI-1Predicted

9.3 / 100

ARC-AGI-1 Cost per TaskPredicted

0.0

ARC-AGI-2Predicted

0.4 / 100

ARC-AGI-2 Cost per TaskPredicted

0.1

BigCodeBench CompletePredicted

56.9 / 100

BigCodeBench InstructPredicted

46.6 / 100

AA Intelligence Index (Matrix)Predicted

69.4

AA Long Context Reasoning (Matrix)Predicted

71.7

AIME 2024Predicted

95.2

AIME 2025 (Matrix)Predicted

95.3

Arena-Hard AutoPredicted

83.7

BrowseCompPredicted

77.3

BRUMO 2025Predicted

90.3

CMIMC 2025Predicted

78.2

CritPtPredicted

0.4

GPQA Diamond (Matrix)Predicted

82.9

GSM8KPredicted

94.8

HLE (Matrix)Predicted

19.4

HMMT Feb 2025Predicted

76.8

HMMT Nov 2025Predicted

60.4

HumanEvalPredicted

82.0

IFBench (Matrix)Predicted

53.3

IFEvalPredicted

77.8

IMO 2025Predicted

18.3

LiveCodeBench (Matrix)Predicted

51.8

MATH-500 (Matrix)Predicted

77.6

MathArena Apex 2025Predicted

1.0

MMLUPredicted

87.7

MMLU-Pro (Matrix)Predicted

84.2

MMMU-ProPredicted

81.8

MRCR v2Predicted

80.7

OSWorldPredicted

40.1

SimpleQAPredicted

33.6

SMT 2025Predicted

83.5

SWE-bench ProPredicted

34.6

Tau-Bench Telecom (Matrix)Predicted

96.1

Terminal-Bench 2.0Predicted

23.6

Terminal-Bench 1.0Predicted

23.6

USAMO 2025Predicted

13.3

Video-MMUPredicted

86.7

browsecompPredicted

56.2

BullshitBenchPredicted

19.8 / 100

Apex AgentsPredicted

2.2

Arc Agi 2Predicted

0.0

BALROGPredicted

0.0

BIG-Bench HardPredicted

3.0

BoolQPredicted

0.7

CAD-EvalPredicted

3.5

Chess PuzzlesPredicted

0.1

CyBenchPredicted

0.3

DeepResearchBenchPredicted

50.0

GdpvalPredicted

0.1

GeoBenchPredicted

0.0

GSM8K (Epoch)Predicted

5.1

GSOPredicted

0.0

HellaSwagPredicted

0.5

HlePredicted

0.1

Lech Mazur WritingPredicted

8.0

METR Time HorizonsPredicted

12.8

OTIS Mock AIME 2024–2025Predicted

0.6

PIQAPredicted

0.8

PosttrainbenchPredicted

0.0

SimpleQA Verified (Epoch)Predicted

0.3

TriviaQAPredicted

6.0

VPCTPredicted

0.4

WinoGrandePredicted

0.9

FrontierMathPredicted

14.6 / 100

GAIA Level 1Predicted

75.6

GAIA Level 2Predicted

65.0

GAIA Level 3Predicted

61.8

GAIAPredicted

60.2 / 100

LegalBenchPredicted

89.7 / 100

LongBench v2 EasyPredicted

55.0

LongBench v2 HardPredicted

48.1

LongBench v2Predicted

43.7 / 100

MATH-500Predicted

1.0 / 100

MathVistaPredicted

70.7 / 100

MLE-benchPredicted

28.0 / 100

MMMUPredicted

78.9 / 100

MMTU Table UnderstandingPredicted

61.0 / 100

MT-BenchPredicted

7.9 / 10

NoLiMa (NIAH)Predicted

94.7 / 100

OCRBench v2Predicted

85.8 / 100

Open LLM AveragePredicted

28.2 / 100

Open LLM: BBHPredicted

58.3 / 100

Open LLM: GPQAPredicted

32.5 / 100

Open LLM: IFEvalPredicted

69.8 / 100

Open LLM: MATH Level 5Predicted

20.4 / 100

Open LLM: MMLU-PROPredicted

42.5 / 100

Open LLM: MUSRPredicted

43.3 / 100

RE-BenchPredicted

16.9

SimpleBenchPredicted

41.8 / 100

simpleqaPredicted

34.0

SWE-bench LitePredicted

40.4 / 100

SWE-bench VerifiedPredicted

61.0 / 100

tau-bench RetailPredicted

79.3 / 100

WebArenaPredicted

5.0 / 100

WeirdMLPredicted

44.7 / 100

WildBenchPredicted

45.7

BFCL (Berkeley Function Calling)bfcl

56.7

Aider Polyglotepoch_ai

0.0

Epoch Capabilities Indexepoch_ai

145.4

FictionLiveBenchepoch_ai

0.5

The Agent Companyepoch_ai

0.4

WebDev Arenaepoch_ai

5162.0

LiveBench Codinglivebench

73.2 / 100

LiveBench Data Analysislivebench

44.3 / 100

LiveBench Languagelivebench

65.6 / 100

LiveBench Mathlivebench

64.4 / 100

LiveBench Overalllivebench

49.9 / 100

LiveBench Reasoninglivebench

45.5 / 100

Vectara Factual Consistencyvectara_hallucination

94.7 / 100

Vectara Hallucination Ratevectara_hallucination

5.3 / 100