Urambazaji
AI BENCHY
Advertise here

AI BENCHY Compare

DeepSeek: DeepSeek V3.2 vs DeepSeek: DeepSeek V4 Pro

Benchmark zimetengenezwa kutoka seti za majaribio za AI BENCHY tarehe: 2026-05-22

Kipimo DeepSeek V3.2 DeepSeek V3.2 medium Toleo: 2025-12-01 DeepSeek V4 Pro DeepSeek V4 Pro high Toleo: 2026-04-24
Alama 7.0 6.6
Nafasi #71 #80
Uaminifu 9.1 9.0
Uthabiti 7.6 8.3
Majaribio sahihi
Kiwango cha kupita kwa kila jaribio 69.2% 66.7%
Majaribio yasiyo thabiti 6 4
Jumla ya uendeshaji 60 60
Gharama kwa matokeo 0.334 1.927
Jumla ya gharama $0.037 $0.212
Bei ya ingizo $0.252 / 1M $0.435 / 1M
Bei ya toleo $0.378 / 1M $0.870 / 1M
Tokeni za matokeo 7,049 12,211
Tokeni za hoja 68,203 53,774
Muda wa majibu (wastani) 53.21s 58.93s
Muda wa majibu (upeo) 189.03s 358.35s
Muda wa majibu (jumla) 1064.26s 1119.75s

Modeli bora kwa alama

Alama dhidi ya gharama ya jumla

Muda wa majibu (wastani)

Alama vs Muda wa majibu (wastani)

Jumla ya tokeni za matokeo

Alama vs Jumla ya tokeni za matokeo

Mgawanyo wa kategoria

Mbinu za kupinga AI Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 9.2 10.0 100.0% 0 24.23s 3,247 6,953
DeepSeek V4 Pro 7.4 10.0 75.0% 0 16.53s 71 3,617
Uandishi wa msimbo Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 3.9 5.8 33.3% 1 184.97s 640 21,230
DeepSeek V4 Pro 2.8 5.0 25.0% 1 51.77s 105 2,641
Mchanganyiko Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 10.0 10.0 100.0% 0 93.11s 571 6,296
DeepSeek V4 Pro 10.0 10.0 100.0% 0 65.02s 465 5,914
Uchanganuzi na uchimbaji wa data Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 10.0 10.0 100.0% 0 36.09s 207 7,693
DeepSeek V4 Pro 8.8 10.0 100.0% 0 23.62s 229 1,710
Mahususi kwa domeni Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 2.9 4.4 22.2% 2 24.27s 21 6,838
DeepSeek V4 Pro 3.0 6.9 16.7% 1 205.66s 10,529 28,089
Akili ya jumla Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 3.8 2.5 50.0% 1 58.29s 49 2,189
DeepSeek V4 Pro 6.1 3.1 66.7% 1 25.09s 76 1,152
Ufuataji wa maagizo Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 10.0 10.0 100.0% 0 35.78s 1,397 2,845
DeepSeek V4 Pro 10.0 10.0 100.0% 0 41.16s 205 2,416
Utatuzi wa mafumbo Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 6.7 5.0 66.7% 2 36.87s 390 6,281
DeepSeek V4 Pro 7.4 7.2 88.9% 1 34.92s 106 3,835
Mwito wa zana Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 10.0 10.0 100.0% 0 34.81s 507 859
DeepSeek V4 Pro 10.0 10.0 100.0% 0 21.33s 372 593
Maarifa ya jumla Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 3.0 10.0 0.0% 0 83.99s 20 7,019
DeepSeek V4 Pro 3.0 10.0 0.0% 0 39.14s 53 3,807

Ulinganisho wa haraka

Badilisha jozi ya ulinganisho