AI BENCHY Compare

OpenAI: GPT-5.4 Mini vs Qwen: Qwen3.5-9B

Benchmark zimetengenezwa kutoka seti za majaribio za AI BENCHY tarehe: 2026-03-17

Kipimo	GPT-5.4 Mini GPT-5.4 Mini none Toleo: 2026-03-17	Qwen3.5-9B Qwen3.5-9B medium Toleo: 2026-03-02

Kipimo	GPT-5.4 Mini GPT-5.4 Mini none Toleo: 2026-03-17	Qwen3.5-9B Qwen3.5-9B medium Toleo: 2026-03-02
Nafasi	#66	#70
Alama	4.8	4.5
Uthabiti	8.6	7.0
Gharama kwa matokeo	0.737	0.787
Jumla ya gharama	$0.030	$0.024
Majaribio sahihi
Kiwango cha kupita kwa kila jaribio	31.4%	35.3%
Majaribio yasiyo thabiti	3	6
Jumla ya uendeshaji	51	51
Tokeni za matokeo	2,085	18,141
Tokeni za hoja	0	141,204
Muda wa majibu (wastani)	1.17s	69.21s
Muda wa majibu (upeo)	2.52s	226.38s
Muda wa majibu (jumla)	19.82s	968.99s

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
GPT-5.4 Mini	3.1	8.1	8.3%	1		929ms	654	0
Qwen3.5-9B	5.1	5.8	50.0%	2		34.44s	2,621	12,411

Mchanganyiko	Alama	Uthabiti	Kiwango cha kupita kwa kila jaribio	Majaribio yasiyo thabiti	Majaribio sahihi	Muda wa majibu (wastani)	Tokeni za matokeo	Tokeni za hoja
GPT-5.4 Mini	3.0	10.0	0.0%	0		2.52s	298	0
Qwen3.5-9B	3.0	10.0	0.0%	0		0ms	0	0

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
GPT-5.4 Mini	10.0	10.0	100.0%	0		1.30s	222	0
Qwen3.5-9B	3.6	5.6	33.3%	1		87.31s	1,383	32,113

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
GPT-5.4 Mini	3.5	4.4	33.3%	2		937ms	88	0
Qwen3.5-9B	3.6	7.2	22.2%	1		137.75s	11,549	48,475

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
GPT-5.4 Mini	4.8	10.0	0.0%	0		1.82s	174	0
Qwen3.5-9B	2.8	1.6	33.3%	1		226.38s	0	30,695

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
GPT-5.4 Mini	6.3	10.0	50.0%	0		728ms	101	0
Qwen3.5-9B	6.4	5.8	66.7%	1		17.15s	599	4,517

Puzzle Solving	Alama	Uthabiti	Kiwango cha kupita kwa kila jaribio	Majaribio yasiyo thabiti	Majaribio sahihi	Muda wa majibu (wastani)	Tokeni za matokeo	Tokeni za hoja
GPT-5.4 Mini	5.4	10.0	33.3%	0		860ms	293	0
Qwen3.5-9B	3.1	10.0	0.0%	0		33.38s	1,545	11,844

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
GPT-5.4 Mini	3.0	10.0	0.0%	0		2.32s	255	0
Qwen3.5-9B	10.0	10.0	100.0%	0		4.31s	444	1,149

Ulinganisho wa haraka

Badilisha jozi ya ulinganisho

Qwen3.5-9BmediumvsGrok 4.1 Fastnone Qwen3.5-9BmediumvsMiMo-V2-Flashnone GPT-5.4 MininonevsQwen3 Coder Nextmedium GPT-5.4 MininonevsGLM 4.7 Flashmedium GPT-5.4 NanononevsQwen3.5-9Bmedium Mercury 2nonevsQwen3.5-9Bmedium GPT-4o-mininonevsQwen3.5-9Bmedium Nemotron 3 Super 120b A12bnoneInapatikana burevsQwen3.5-9Bmedium Mistral Small 4mediumvsGPT-5.4 Mininone Mistral Small 4nonevsQwen3.5-9Bmedium Trinity Large PreviewnoneInapatikana burevsQwen3.5-9Bmedium Kimi K2.5nonevsQwen3.5-9Bmedium