Model yang Dibandingkan

Perbandingan benchmark Qwen3.8 27B (low) vs Qwen3.8 Flash Next (low) vs Qwen3.8 27B (high) vs Qwen3.8 Flash Next (xhigh): Qwen3.8 27B (high) unggul pada Skor dengan 8.7. Qwen3.8 27B (low) unggul pada Keandalan dengan 10.0. Qwen3.8 Flash Next (low) memiliki Total Biaya terendah di ~$0.045. Qwen3.8 Flash Next (low) paling cepat di 23.66s.

Benchmark dihasilkan dari suite pengujian AI BENCHY pada: 2026-10-05

Model yang Dibandingkan

Peringkat: #121
Total token output: 209,143
Waktu respons (rata-rata): 46.60s
Total Biaya: ~$0.089

Peringkat: #96
Total token output: 135,295
Waktu respons (rata-rata): 23.66s
Total Biaya: ~$0.045

Peringkat: #55
Total token output: 677,080
Waktu respons (rata-rata): 166.34s
Total Biaya: ~$0.298

Peringkat: #154
Total token output: 621,056
Waktu respons (rata-rata): 102.66s
Total Biaya: ~$0.195

Model yang direkomendasikan Qwen3.8 Flash Next (low)

Its score stays close to the best score here (8.0 vs 8.7), while costing about 4.3x less than model lain dalam perbandingan ini.

Perbandingan terperinci

Metrik	Qwen3.8 27B Qwen3.8 27B low Rilis: 2026-08-14 Tersedia gratis	Qwen3.8 Flash Next Qwen3.8 Flash Next low Rilis: 2026-10-04	Qwen3.8 27B Qwen3.8 27B high Rilis: 2026-08-14 Tersedia gratis	Qwen3.8 Flash Next Qwen3.8 Flash Next xhigh Rilis: 2026-10-04

Metrik	Qwen3.8 27B Qwen3.8 27B low Rilis: 2026-08-14 Tersedia gratis	Qwen3.8 Flash Next Qwen3.8 Flash Next low Rilis: 2026-10-04	Qwen3.8 27B Qwen3.8 27B high Rilis: 2026-08-14 Tersedia gratis	Qwen3.8 Flash Next Qwen3.8 Flash Next xhigh Rilis: 2026-10-04
Skor	7.6	8.0	8.7	7.1
Peringkat	#121	#96	#55	#154
Keandalan	10.0	10.0	9.7	9.9
Konsistensi	9.7	8.0	9.2	7.9
Percobaan	69/69	69/69	69/69	69/69
Tes benar
Tingkat lulus per percobaan	68.1%	79.7%	78.3%	72.5%
Tes tidak stabil	1	6	2	6
Total Run	69	69	69	69
Biaya per hasil	~0.589	~0.299	~1.751	~1.390
Total Biaya	~$0.089	~$0.045	~$0.298	~$0.195
Harga input	T/A	T/A	T/A	T/A
Harga output	T/A	T/A	T/A	T/A
Harga baca cache	T/A	T/A	T/A	T/A
Harga tulis cache	T/A	T/A	T/A	T/A
Total token input	315,290	342,163	348,967	301,749
Token output	1,607	1,518	1,005	1,033
Token penalaran	207,536	133,777	676,075	620,023
Waktu respons (rata-rata)	46.60s	23.66s	166.34s	102.66s
Waktu respons (maks)	376.04s	166.72s	640.85s	674.08s
Waktu respons (total)	1071.83s	544.16s	3825.81s	2361.10s
Parameter	27.3B	180B total (6B aktif)	27.3B	180B total (6B aktif)
Ketersediaan	Bobot tersedia	Bobot tersedia	Bobot tersedia	Bobot tersedia

Harga cache berlaku untuk token input. Pembacaan memakai kembali prompt tersimpan; penulisan menyimpannya dan dapat menambah biaya. Token output menggunakan harga output.

Showcase generasi model

Hamster playing table tennis

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

#121 Qwen3.8 27B

low

Biaya: ~$0.003
Waktu: 92.8s
Token: 6,902 tok

#96 Qwen3.8 Flash Next

low

Biaya: ~$0.002
Waktu: 47.2s
Token: 5,270 tok

#55 Qwen3.8 27B

high

Biaya: ~$0.008
Waktu: 270.1s
Token: 18,963 tok

#154 Qwen3.8 Flash Next

xhigh

Biaya: ~$0.010
Waktu: 333.1s
Token: 30,654 tok

Model teratas berdasarkan skor

Skor vs Total Biaya

Waktu respons (rata-rata)

Skor vs Waktu respons (rata-rata)

Total token output

Skor vs Total token output

Rincian Kategori

Kategori:

Tugas agen	Skor	Konsistensi	Tingkat lulus per percobaan	Tes tidak stabil	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	6.1	3.1	66.7%	1	211.32s	215,585	62	39,752
Qwen3.8 Flash Next	6.1	3.1	66.7%	1	137.50s	194,811	72	28,519
Qwen3.8 27B	10.0	10.0	100.0%	0	132.27s	248,788	101	20,344
Qwen3.8 Flash Next	4.7	3.1	33.3%	1	173.66s	148,808	26	38,616

Trik anti-AI	Skor	Konsistensi	Tingkat lulus per percobaan	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	10.0	10.0	100.0%	3.02s	984	183	2,016
Qwen3.8 Flash Next	10.0	10.0	100.0%	2.03s	984	233	1,773
Qwen3.8 27B	10.0	10.0	100.0%	8.10s	1,128	114	5,319
Qwen3.8 Flash Next	10.0	10.0	100.0%	6.73s	1,128	184	7,283

Pemrograman	Skor	Konsistensi	Tingkat lulus per percobaan	Tes tidak stabil	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	7.7	10.0	66.7%	0	140.92s	8,127	585	86,694
Qwen3.8 Flash Next	10.0	10.0	100.0%	0	29.46s	8,127	397	24,147
Qwen3.8 27B	8.1	7.0	88.9%	1	248.62s	8,235	346	143,335
Qwen3.8 Flash Next	7.4	7.0	77.8%	1	192.20s	8,235	122	186,070

Gabungan	Skor	Konsistensi	Tingkat lulus per percobaan	Tes tidak stabil	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	10.0	10.0	100.0%	0	90.63s	69,687	66	37,791
Qwen3.8 Flash Next	6.9	5.9	66.7%	1	98.91s	118,012	60	51,088
Qwen3.8 27B	10.0	10.0	100.0%	0	185.57s	68,997	66	72,556
Qwen3.8 Flash Next	7.3	5.8	83.3%	1	92.24s	122,809	52	43,355

Parsing dan ekstraksi data	Skor	Konsistensi	Tingkat lulus per percobaan	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	10.0	10.0	100.0%	8.03s	9,087	150	1,962
Qwen3.8 Flash Next	10.0	10.0	100.0%	6.20s	8,253	267	2,783
Qwen3.8 27B	10.0	10.0	100.0%	34.85s	9,675	123	10,320
Qwen3.8 Flash Next	10.0	10.0	100.0%	9.24s	8,325	212	3,925

Spesifik domain	Skor	Konsistensi	Tingkat lulus per percobaan	Tes tidak stabil	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	5.3	10.0	33.3%	0	60.78s	1,014	12	31,872
Qwen3.8 Flash Next	4.1	4.4	44.5%	2	23.34s	1,014	12	18,966
Qwen3.8 27B	5.3	10.0	33.3%	0	526.74s	975	6	234,555
Qwen3.8 Flash Next	3.5	4.4	33.3%	2	326.96s	1,122	11	239,950

Kecerdasan umum	Skor	Konsistensi	Tingkat lulus per percobaan	Tes tidak stabil	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	5.0	10.0	0.0%	0	5.37s	594	132	729
Qwen3.8 Flash Next	4.7	3.1	33.3%	1	5.00s	594	130	1,171
Qwen3.8 27B	5.0	10.0	0.0%	0	8.45s	630	63	1,191
Qwen3.8 Flash Next	4.7	3.1	33.3%	1	3.95s	630	82	864

Kepatuhan instruksi	Skor	Konsistensi	Tingkat lulus per percobaan	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	10.0	10.0	100.0%	2.43s	855	48	915
Qwen3.8 Flash Next	9.8	10.0	100.0%	2.32s	855	48	1,050
Qwen3.8 27B	9.8	10.0	100.0%	8.24s	927	42	2,871
Qwen3.8 Flash Next	9.8	10.0	100.0%	3.72s	927	42	1,939

Pemecahan teka-teki	Skor	Konsistensi	Tingkat lulus per percobaan	Tes tidak stabil	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	7.7	10.0	66.7%	0	4.91s	930	240	2,382
Qwen3.8 Flash Next	8.7	7.8	88.9%	1	1.72s	930	272	1,001
Qwen3.8 27B	7.6	7.2	77.8%	1	199.90s	1,038	120	138,427
Qwen3.8 Flash Next	7.7	10.0	66.7%	0	19.51s	1,038	275	15,050

Pemanggilan alat	Skor	Konsistensi	Tingkat lulus per percobaan	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	3.0	10.0	0.0%	8.42s	8,145	15	1,350
Qwen3.8 Flash Next	10.0	10.0	100.0%	5.85s	8,301	21	816
Qwen3.8 27B	10.0	10.0	100.0%	5.62s	8,256	21	912
Qwen3.8 Flash Next	10.0	10.0	100.0%	5.11s	8,409	21	650

Pengetahuan umum	Skor	Konsistensi	Tingkat lulus per percobaan	Waktu respons (rata-rata)	Token input	Token output	Token penalaran
Qwen3.8 27B	3.0	10.0	0.0%	12.64s	282	114	2,073
Qwen3.8 Flash Next	3.0	10.0	0.0%	9.26s	282	6	2,463
Qwen3.8 27B	3.0	10.0	0.0%	263.98s	318	3	46,245
Qwen3.8 Flash Next	3.0	10.0	0.0%	325.04s	318	6	82,321

Perbandingan Cepat

Ganti Pasangan Perbandingan

GPT-5.2mediumvsQwen3.8 27BhighTersedia gratis Qwen3.8 27BhighTersedia gratisvsMiMo-V2.6-Promedium Seed 2.1 TurbomediumvsQwen3.8 27BhighTersedia gratis Gemini 3.8 FlashmediumvsQwen3.8 Flash Nextlow GPT-5.4 NanomediumvsQwen3.8 Flash Nextlow GPT-6 SollowvsQwen3.8 27BhighTersedia gratis Qwen3.8 27BlowTersedia gratisvsSpace Bunny Alphahigh Qwen3.8 27BlowTersedia gratisvsGLM 5.1medium Qwen3.8 Flash NextlowvsGLM 5.3 FlashXmax Qwen3.8 Flash NextxhighvsGLM 5.3 FlashXmedium Claude Fable 5.1highvsQwen3.8 Flash Nextlow Nemotron 3 UltramediumTersedia gratisvsQwen3.8 Flash Nextlow