Urambazaji
AI BENCHY
Advertise here

DeepSeek V3.2 vs Granite 4.2 8B (medium)

Granite 4.2 8B (medium) inaongoza kwa average score: 5.2 vs 5.0. Granite 4.2 8B (medium) ina gharama ya chini ya benchmark: $0.009 vs $0.054. DeepSeek V3.2 ni ya haraka zaidi: 17.89s vs 45.97s, na pass rates 36.4% vs 34.9%.

Benchmark zimetengenezwa kutoka seti za majaribio za AI BENCHY tarehe: 2026-09-02

Nafasi
#256
Jumla ya tokeni za matokeo
42,099
Muda wa majibu (wastani)
17.89s
Jumla ya gharama
$0.054
Nafasi
#243
Jumla ya tokeni za matokeo
19,330
Muda wa majibu (wastani)
45.97s
Jumla ya gharama
$0.009
Muundo unaopendekezwa Granite 4.2 8B (medium)

It has the best score here (5.2), while costing about 6.5x less than DeepSeek V3.2.

Ulinganisho wa kina

Kipimo DeepSeek V3.2 DeepSeek V3.2 none Toleo: 2025-12-01 Granite 4.2 8B Granite 4.2 8B medium Toleo: 2026-09-02
Alama 5.0 5.2
Nafasi #256 #243
Uaminifu 10.0 8.7
Uthabiti 8.1 9.6
Majaribio 66/66 66/66
Majaribio sahihi
Kiwango cha kupita kwa kila jaribio 36.4% 34.9%
Majaribio yasiyo thabiti 5 1
Jumla ya uendeshaji 66 66
Gharama kwa matokeo 0.870 0.117
Jumla ya gharama $0.054 $0.009
Bei ya ingizo $0.269 / 1M $0.100 / 1M
Bei ya toleo $0.400 / 1M $0.150 / 1M
Jumla ya tokeni za ingizo 135,831 52,909
Tokeni za matokeo 42,099 7,058
Tokeni za hoja 0 12,272
Muda wa majibu (wastani) 17.89s 45.97s
Muda wa majibu (upeo) 115.89s 557.96s
Muda wa majibu (jumla) 393.53s 1011.42s
Vigezo 671B jumla (37B vinavyotumika) 8B
Upatikanaji Chanzo huria Chanzo huria

Onyesho la kizazi cha modeli

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#256 DeepSeek V3.2

none
Gharama
$0.002
Muda
7.0s
Tokeni
1,046 tok

#243 IBM: Granite 4.2 8B

medium
Provider returned error
Gharama
$0.000
Muda
0.2s
Tokeni
0 tok

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

Uandishi wa msimbo Alama Uthabiti Kiwango cha kupita kwa kila jaribio Majaribio yasiyo thabiti Majaribio sahihi Muda wa majibu (wastani) Tokeni za ingizo Tokeni za matokeo Tokeni za hoja
DeepSeek V3.2 3.1 6.9 11.1% 1 14.54s 7,279 4,528 0
Granite 4.2 8B 5.5 10.0 33.3% 0 12.09s 8,439 2,445 3,900

Ulinganisho wa haraka

Badilisha jozi ya ulinganisho