DeepSeek V3 (Dec '24)

Last updated: Recently

Model Type

Text

Max Context

164K

Max Output

163,840 tokens

Parameter Scale

DeepSeek V3.2 Speciale is high compute variant of DeepSeek V3.2 optimized for maximum reasoning and agentic performance. 163,840 token context window, maximum output of 163,840 tokens.

Input Modality

text

Output Modality

text

Inference Speed

Latest Release Date

12/26/2024

SDK Ecosystem

artificial-analysismanufactureraa-bootstrap

Model Overview Fields

name

DeepSeek V3 (Dec '24)

Release date

12/26/2024

Performance

model_creator.name

DeepSeek

AA Evaluation Scores

AA Intelligence Index

16.5

AA Coding Index

16.4

AA Math Index

MMLU Pro

0.752

GPQA

0.557

HLE

0.036

LiveCodeBench

0.359

SciCode

0.354

Math-500

0.886666666666667

AIME

0.253333333333333

AIME 2025

0.26

IFBench

0.347619047619048

LCR

0.29

TerminalBench Hard

0.0681818181818182

TAU2

0.228070175438596

Output Speed (tokens/s)

TTFT (s)

First Answer Token (s)

DeepSeek

DeepSeek V3 (Dec '24)

$0.40/M$0.89/M

DeepSeek V3.2 Speciale is high compute variant of DeepSeek V3.2 optimized for maximum reasoning and agentic performance. 163,840 token context window, maximum output of 163,840 tokens.

There are no reviews for this model yet.

artificial-analysismanufactureraa-bootstrap

0.0