AI BENCHY Compare

Google: Gemma 4 31B vs Qwen: Qwen3.5-27B

Benchmark zimetengenezwa kutoka seti za majaribio za AI BENCHY tarehe: 2026-05-19

Kipimo	Gemma 4 31B Gemma 4 31B medium Toleo: 2026-04-02 Inapatikana bure	Qwen3.5-27B Qwen3.5-27B medium Toleo: 2026-02-24

Kipimo	Gemma 4 31B Gemma 4 31B medium Toleo: 2026-04-02 Inapatikana bure	Qwen3.5-27B Qwen3.5-27B medium Toleo: 2026-02-24
Alama	8.2	8.1
Nafasi	#18	#22
Uaminifu	6.7	10.0
Uthabiti	9.6	8.8
Majaribio sahihi
Kiwango cha kupita kwa kila jaribio	77.2%	77.2%
Majaribio yasiyo thabiti	1	3
Jumla ya uendeshaji	57	57
Gharama kwa matokeo	0.158	4.107
Jumla ya gharama	$0.023	$0.534
Bei ya ingizo	$0.120 / 1M	$0.195 / 1M
Bei ya toleo	$0.370 / 1M	$1.560 / 1M
Tokeni za matokeo	14,426	2,531
Tokeni za hoja	37,964	266,183
Muda wa majibu (wastani)	28.72s	54.71s
Muda wa majibu (upeo)	90.14s	163.96s
Muda wa majibu (jumla)	488.27s	1039.57s

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
Gemma 4 31B	10.0	10.0	100.0%	0		12.89s	962	2,046
Qwen3.5-27B	8.7	7.9	91.7%	1		19.75s	569	31,505

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
Gemma 4 31B	4.7	1.6	66.7%	1		70.97s	3,166	5,449
Qwen3.5-27B	10.0	10.0	100.0%	0		70.35s	375	19,165

Mchanganyiko	Alama	Uthabiti	Kiwango cha kupita kwa kila jaribio	Majaribio yasiyo thabiti	Majaribio sahihi	Muda wa majibu (wastani)	Tokeni za matokeo	Tokeni za hoja
Gemma 4 31B	3.0	10.0	0.0%	0		0ms	0	0
Qwen3.5-27B	10.0	10.0	100.0%	0		163.96s	483	9,991

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
Gemma 4 31B	10.0	10.0	100.0%	0		21.11s	1,822	2,951
Qwen3.5-27B	10.0	10.0	100.0%	0		30.26s	270	16,150

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
Gemma 4 31B	7.7	10.0	66.7%	0		38.48s	4,349	8,985
Qwen3.5-27B	5.3	10.0	33.3%	0		79.53s	43	52,368

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
Gemma 4 31B	10.0	10.0	100.0%	0		9.57s	105	888
Qwen3.5-27B	6.1	3.1	66.7%	1		101.41s	70	23,147

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
Gemma 4 31B	10.0	10.0	100.0%	0		12.76s	533	2,035
Qwen3.5-27B	10.0	10.0	100.0%	0		19.66s	97	11,638

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
Gemma 4 31B	9.9	10.0	100.0%	0		27.63s	1,797	5,596
Qwen3.5-27B	8.2	7.7	77.8%	1		64.61s	245	77,213

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
Gemma 4 31B	3.0	10.0	0.0%	0		0ms	0	0
Qwen3.5-27B	10.0	10.0	100.0%	0		7.45s	348	1,323

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
Gemma 4 31B	3.0	10.0	0.0%	0		90.14s	1,692	10,014
Qwen3.5-27B	3.0	10.0	0.0%	0		85.11s	31	23,683

Ulinganisho wa haraka

Badilisha jozi ya ulinganisho

Gemini 3.5 FlashminimalvsQwen3.5-27Bmedium Gemini 3 Flash PreviewnonevsQwen3.5-27Bmedium Gemini 3.1 Flash Lite PreviewlowvsQwen3.5-27Bmedium Gemini 3.1 Flash Lite PreviewnonevsQwen3.5-27Bmedium GPT-5.2 ChatnonevsQwen3.5-27Bmedium Gemini 3.1 Flash LitelowvsQwen3.5-27Bmedium GPT-5.3 ChatnonevsQwen3.5-27Bmedium DeepSeek V4 FlashhighInapatikana burevsQwen3.5-27Bmedium Gemma 4 31BmediumInapatikana burevsGPT-5.2 Chatnone Gemma 4 31BmediumInapatikana burevsGPT-5.5low Gemma 4 31BmediumInapatikana burevsGPT-5.3 Chatnone DeepSeek V4 FlashhighInapatikana burevsGemma 4 31BmediumInapatikana bure