Saltar al contenido principal

DeepSeek V3.1 (Non-reasoning)

DeepSeekDeepSeekOpen WeightMIT · Uso Comercial

Descripción

DeepSeek-V3.1 is a hybrid model supporting both thinking and non-thinking modes through different chat templates. Built on DeepSeek-V3.1-Base with a two-phase long context extension (32K phase: 630B tokens, 128K phase: 209B tokens), it features 671B total parameters with 37B activated. Key improvements include smarter tool calling through post-training optimization, higher thinking efficiency achieving comparable quality to DeepSeek-R1-0528 while responding more quickly, and UE8M0 FP8 scale data format for model weights and activations. The model excels in both reasoning tasks (thinking mode) and practical applications (non-thinking mode), with particularly strong performance in code agent tasks, math competitions, and search-based problem solving.

Fecha de lanzamiento
2025-08-21
Parámetros
671.0B
Longitud del contexto
164K
Modalidades
text

Radar de capacidades

60
general
60
coding
60
reasoning
60
scienceest.
30
agents
0
multimodal

Science se estima a partir de las puntuaciones científicas de LLM Stats o del razonamiento cuando no hay benchmarks científicos dedicados.

Rankings

No hay datos de ranking disponibles

Puntuaciones de benchmarks (LLM Stats)

(LLM Stats (zeroeval))

Factuality

SimpleQA93.4%Aut.

General

Aider-Polyglot68.4%Aut.

Language

MMLU-Redux91.8%Aut.
MMLU-Pro83.7%Aut.

Math

CodeForces0.70 / 3000Aut.
AIME 202466.3%Aut.
AIME 202549.8%Aut.
HMMT 202533.5%Aut.

Reasoning

GPQANYU + Cohere + Anthropic (2023)74.9%Aut.
SWE-Bench Verified66.0%Aut.
LiveCodeBench56.4%Aut.
SWE-bench Multilingual54.5%Aut.
BrowseComp-zh49.2%Aut.
Terminal-Bench31.3%Aut.
BrowseCompOpenAI (2025)30.0%Aut.
Humanity's Last Exam15.9%Aut.

Índices de evaluación AA

(Artificial Analysis)

No hay datos de evaluación AA disponibles

Puntuaciones por categoría LLM Stats

(LLM Stats (zeroeval))
Language
90
Factuality
90
Legal
80
Finance
80
Healthcare
80
Physics
70
Frontend Development
70
Biology
70
Chemistry
70
Math
60
Reasoning
60
General
60
Code
60
Search
40
Agents
30
Vision
20

Precios

Precio de entrada$0.53 / 1M tokens
Precio de salida$1.59 / 1M tokens
Precio mixto (3:1)$0.425 / 1M tokens
Precio de lectura caché$0.13 / 1M tokens

Velocidad

Tokens/seg91.4
Retraso del primer token1.29s
Tiempo hasta la respuesta1.29s

Ranking de Precios por Proveedor

Ranking de Precios por Proveedor

5 proveedores

Más barato: DeepInfraMás caro: Alibaba (China)
ProveedorEntradaSalida
1DeepInfraMás barato
$0.25
$0.95
2OpenRouter
$0.25
$0.95
3Kilo Gateway
$0.25
$0.95
4DeepSeekPRINCIPAL
$0.53
$1.59
5Alibaba (China)
$0.574
$1.721

Comparar precios entre diferentes proveedores de API para este modelo.

Fuentes externas