DeepSeek V3.2 (Thinking)

Release Date: 12/1/2025

Benchmarked by Vals Logo
Vals Index

Accuracy (Vals Index)

37.28%

Latency (Vals Index)

922.78s

Cost/Test (Vals Index)

$0.11

Context Window

160k

Max Output Tokens

20k

Input Modality

Hyperparameter settings
Default Provider : Fireworks AI

Temperature

1

Top P

Default

Top K

Default

Max Output Tokens

20,480

Reasoning Effort

high

Benchmarks

Accuracy

Rankings

Vals Index

0.0%

27/ 31

CaseLaw (v2)

0.0%

60/ 66

CorpFin

0.0%

58/ 85

ProofBench

0.0%

16/ 17

TaxEval (v2)

0.0%

57/ 93

AIME

0.0%

29/ 85

GPQA

0.0%

23/ 88

IOI

0.0%

20/ 45

LiveCodeBench

0.0%

26/ 95

LegalBench

0.0%

61/ 107

MedQA

0.0%

18/ 93

MMLU Pro

0.0%

20/ 85

SWE-bench

0.0%

23/ 55

Academic Benchmarks
Proprietary Benchmarks (contact us to get access)

Join our mailing list to receive benchmark updates

Model benchmarks are seriously lacking. With Vals AI, we report how language models perform on the industry-specific tasks where they will be used.

By subscribing, I agree to Vals' Privacy Policy.