Voltar para Explorar
DeepSeek: DeepSeek V3.1 Terminus
DeepSeek·Lançado em 2025-09-22
Open Source164K ctxMoE
Sobre
DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's...
Modelos Relacionados
DeepSeek: DeepSeek V4 Pro2026-04-24DeepSeek V4 Pro (Reasoning, High Effort)2026-04-24DeepSeek: DeepSeek V4 Flash2026-04-24DeepSeek V4 Flash (Reasoning, High Effort)2026-04-24DeepSeek V4 Pro (Non-reasoning)2026-04-24DeepSeek V4 Flash (Non-reasoning)2026-04-24DeepSeek: DeepSeek V4 Flash (free)2026-04-24DeepSeek V3.2 (Reasoning)2025-12-01
Preços
Entrada
$0.27
por 1M tokens
Saída
$0.95
por 1M tokens
Combinado
$0.44
por 1M tokens
Mais barato que 54% dos modelos. Preço mediano é $0.54/1M tokens.
Calculadora de Custo
Tokens por dia1M
100K100M
Diário
$0.44
Mensal
$13.20
vs. Modelos Similares
DeepSeek V3.2 Exp (Non-reasoning)Q:-0.1
$0.31-30%
Qwen: Qwen3 Coder NextQ:-0.2
$0.28-36%
Grok Code Fast 1Q:+0.2
$0.53+19%
DeepSeek V3.1 (Non-reasoning)Q:-0.4
$0.83+90%
Desempenho
Janela de Contexto
164K
tokens
Maior que 48% dos modelos
Saída Máxima
33K
tokens
20% do contexto
Comparação de Janela de Contexto
DeepSeek V3.2 Speciale
164KIgual
DeepSeek: R1
164KIgual
DeepSeek: R1 0528
164KIgual
Benchmarks
MMLU-Pro
83.6%
GPQA Diamond
75.1%
HLE
8.4%
LiveCodeBench
52.9%
SciCode
32.1%
TerminalBench Hard
31.8%
MATH-500Não avaliado
AIMENão avaliado
AIME 2025
53.7%
IFBench
41.2%
Long Context Recall
43.3%
Tau2
37.1%
Média do MercadoMelhor Score