Watts

Name: OpenWeights Terminal measured pairs
Creator: Dvidia
License: https://owterminal.com/cite

Decode ÷ named TDP. No TDP, no row. Not a composite score.

#	Model · Build	Hardware	Decode · tok/s	TDP	tok/100W	Hours / 1M
1	Qwen3.6-35B-A3B	M4 Pro 64GB	108.6	45W	241.3	2.56
2	Qwen3.8-27B	M6 mini 32GB	61.9	27W	229.3	4.49
3	Qwen3.8-27B	M5 Pro 64GB	89.9	40W	224.8	3.09
4	Qwen3.6-35B-A3B	M4 Pro	85.5	45W	190	3.25
5	Qwen3.5-4B	M4 Pro	82.8	45W	184	3.35
6	Qwen3.8-Flash-Next	M5 Max	126.5	70W	180.7	2.2
7	Qwen3.5-9B	M4 Pro	49.3	45W	109.6	5.63
8	Qwen3.8-27B	RTX PRO 6000 Blackwell Max-Q	272.8	250W	109.1	1.02
9	Qwen3-8B	M4 Pro	48.3	45W	107.3	5.75
10	Qwen3.8-Flash-Next	M5 Pro 64GB	41.4	40W	103.5	6.71
11	AliceAI-Foundation-80B-A3B	M5 Max 128GB	65	70W	92.9	4.27
12	Qwen3.8-27B	RTX A5000	177.6	230W	77.2	1.56
13	Qwen3.8-Flash-Next Coder	M5 Max 64GB	51	70W	72.9	5.45
14	Qwen3.8-27B	RTX 5090	370	575W	64.3	0.75
15	Qwen3.6-35B-A3B	M1 Max 32GB	37	60W	61.7	7.51
16	Qwen3.5 2B	RTX 5090 32GB	351	575W	61	0.79
17	Qwen3.8-27B	RTX PRO 6000 Blackwell Max-Q	151.4	250W	60.6	1.83
18	Qwen3.8-27B	RTX A5000	139.3	230W	60.6	1.99
19	Gemma-4 26B-A4B-it	Radeon RX 9060 XT 16GB	88	160W	55	3.16
20	GLM 5.3 Flash	4× DGX Spark	127	240W	52.9	2.19
21	Swift 1.5 Qwen3.8 Flash-Next	RTX 4090 24GB	220	450W	48.9	1.26
22	Qwen3.8-27B	RTX 5090	258	575W	44.9	1.08
23	GLM 5.3 Flash	4× DGX Spark	104	240W	43.3	2.67
24	DeepSeek v4.1 Flash	2× DGX Spark	103	240W	42.9	2.7
25	Qwen3.8-27B	DGX Spark	102.9	240W	42.9	2.7
26	GLM 5.3 Flash	2× RTX PRO 6000	238	600W	39.7	1.17
27	Qwen3.6-35B-A3B	2× DGX Spark	94.4	240W	39.3	2.94
28	DeepSeek v4.1 Flash	2× DGX Spark	93.6	240W	39	2.97
29	Qwen3.8-Flash-Next	Ryzen AI Max+ 395 128GB	53.8	140W	38.4	5.16
30	GLM 5.3 Flash	2x DGX Spark	85	240W	35.4	3.27
31	Qwen3.8-Flash-Next Coder	RTX 5060 Ti 16GB + RTX 3070 8GB	127	370W	34.3	2.19
32	Qwen3.8-Flash-Next	RTX 5090	195.8	575W	34.1	1.42
33	Nemotron 3.5 Bolt 74B	DGX Spark	80	240W	33.3	3.47
34	Qwen3.8-Flash-Next	RTX 4060 Ti 16GB	55	165W	33.3	5.05
35	Qwen3.8-Flash-Next	RTX 4060 Ti 16GB	55	165W	33.3	5.05
36	Qwen3.8-Flash-Next	RTX 5090	191	575W	33.2	1.45
37	Qwen3.8-Flash-Next	DGX Spark	79.5	240W	33.1	3.49
38	MiMo-V2.6-Distill-Qwen-9B	RTX 5060 8GB	47	145W	32.4	5.91
39	Qwen3.8-27B	RTX 5090	186	575W	32.3	1.49
40	Ornith	GTX 1660 SUPER 6GB	40	125W	32	6.94
41	Qwen3.8-Flash-Next	DGX Spark	74.8	240W	31.2	3.71
42	Ternary Bonsai 2 27B	RTX 5060 Ti 16GB	54.7	180W	30.4	5.08
43	Qwen3.8-Flash-Next	RTX 4060 Ti	50	165W	30.3	5.56
44	Gemma-4 26B-A4B-it	RTX 5090 32GB	173	575W	30.1	1.61
45	Qwen3.8-27B	DGX Spark	72	240W	30	3.86
46	Empero Qwen3.8-35B-A3B	RTX 3060 12GB	50	170W	29.4	5.56
47	Ternary Bonsai 2 27B	RTX 3060 12GB	50	170W	29.4	5.56
48	Qwen3.8-Flash-Next	RTX 3090 24GB	100.6	350W	28.7	2.76
49	Qwen3.8-Flash-Next	RTX 3090 24GB	100	350W	28.6	2.78
50	Qwen3.8-Flash-Next	2× RTX 5060 Ti 16GB	99.9	360W	27.8	2.78
51	Qwen3.8-Flash-Next	DGX Spark	65	240W	27.1	4.27
52	Qwen3.8-Flash-Next	DGX Spark	65	240W	27.1	4.27
53	Qwen3.8-Flash-Next	RTX 5070 12GB	66.7	250W	26.7	4.16
54	GLM 5.3 Flash	4× RTX PRO 6000	264	1000W	26.4	1.05
55	GLM 5.3 Flash	2× DGX Spark	60.4	240W	25.2	4.6
56	Qwen3.8-Flash-Next	RTX 3090 24GB	87	350W	24.9	3.19
57	Qwen3.8-Flash-Next	RTX 5090 32GB	142	575W	24.7	1.96
58	Qwen3.8-Flash-Next	2× DGX Spark	58	240W	24.2	4.79
59	GLM 5.3 Flash	2x DGX Spark	55.6	240W	23.2	5
60	Qwen3.8-Flash-Next	Tesla V100 32GB	57	250W	22.8	4.87
61	CYBER-FROST 3.8	DGX Spark	54.7	240W	22.8	5.08
62	Qwen3.8-Flash-Next	2× RTX 3090 24GB	158.3	700W	22.6	1.75
63	Qwen3.8-Flash-Next	RTX 3080 Ti 12GB	79	350W	22.6	3.52
64	UUnderdog Saluki 27B	DGX Spark	53	240W	22.1	5.24
65	Qwen3.8-Flash-Next	2× RTX 3090 24GB	153.4	700W	21.9	1.81
66	Qwen3.8-Flash-Next	Intel Arc B570 10GB	32.4	150W	21.6	8.57
67	Qwen3.8-Flash-Next	DGX Spark	50.2	240W	20.9	5.54
68	Qwen3.8-Flash-Next	RTX 3090 24GB	72	350W	20.6	3.86
69	Qwen3.8-Flash-Next	RTX 4090	92	450W	20.4	3.02
70	GLM 5.3 Flash	RTX PRO 6000 + 2× GB10	120	600W	20	2.31
71	Qwen3.8-Flash-Next	RTX 3090 24GB	70	350W	20	3.97
72	Qwen3.8-Flash-Next	RTX 5090	114.3	575W	19.9	2.43
73	Qwen3.8-Flash-Next	2× DGX Spark	45	240W	18.8	6.17
74	Qwen3.8-Flash-Next	RTX 5090 32GB	104	575W	18.1	2.67
75	Qwen3.8-Flash-Next	RTX 5070 Ti 16GB	53.5	300W	17.8	5.19
76	Qwen3.8-27B	DGX Spark	42.3	240W	17.6	6.57
77	Qwen3.8-Flash-Next	2× RTX 3090 24GB	122.5	700W	17.5	2.27
78	GLM 5.3	8× DGX Spark	40	240W	16.7	6.94
79	Qwen3.8-Flash-Next	RTX 3080 10GB	52.7	320W	16.5	5.27
80	Qwen3.8-Flash-Next	RTX 3060 12GB	27.7	170W	16.3	10.03
81	GLM 5.3 Flash	Mac Studio M4 Max 64GB	11.3	70W	16.2	24.52
82	GLM 5.3	4x DGX Spark	38.4	240W	16	7.23
83	Qwen3.8-Flash-Next	RTX PRO 6000	92.8	600W	15.5	2.99
84	GLM 5.3	4× DGX Spark	35	240W	14.6	7.94
85	Qwen3.8-Flash-Next	RTX 5090 32GB	80.9	575W	14.1	3.43
86	Qwen3.8-Flash-Next	RTX 5080 16GB	50	360W	13.9	5.56
87	RRED-SNOW 5.3 Flash	DGX Spark	33.0	240W	13.7	8.43
88	GLM 5.3 Flash	DGX Spark	32.3	240W	13.5	8.59
89	Qwen3.8-Flash-Next	DGX Spark	32	240W	13.3	8.68
90	Qwen3.8-27B	DGX Spark	31.8	240W	13.3	8.74
91	Qwen3.8-27B	RTX 5060 Ti 16GB	23	180W	12.8	12.08
92	GLM 5.3	4x DGX Spark	30	240W	12.5	9.26
93	GLM 5.3 Flash	DGX Spark	30	240W	12.5	9.26
94	Qwen3.8-27B	DGX Spark	28.7	240W	12	9.68
95	Qwen3.8-Flash-Next	2× RTX 3060 12GB	40	340W	11.8	6.94
96	Qwen3.8-Flash-Next	M5 Max 64GB	7.8	70W	11.1	35.61
97	Qwen3.8-27B	RTX 5060 Ti 16GB	19	180W	10.6	14.62
98	DeepSeek v4.1 Flash	RTX PRO 6000	57.7	600W	9.6	4.81
99	Qwen3.8-27B	RTX 4090 24GB	40.7	450W	9	6.83
100	Qwen3.8-Flash-Next	AMD MI50 32GB	26.8	300W	8.9	10.36
101	Qwen3.8-Flash-Next	2× AMD MI50 32GB	52	600W	8.7	5.34
102	Nex-N2.5-mini	RTX 5060 Ti 16GB	14	180W	7.8	19.84
103	GLM 5.3 Flash	DGX Spark	18.6	240W	7.7	14.96
104	Qwen3.8-Flash-Next	RTX 5090	42.1	575W	7.3	6.6
105	Qwen3.8-Flash-Next	RTX 3090 + RTX 2070 SUPER	34.2	500W	6.8	8.12
106	GLM 5.3 Flash	2x DGX Spark	15	240W	6.3	18.52
107	Qwen3.8-Flash-Next	4× Tesla V100 32GB	60	1000W	6	4.63
108	Qwen3.8-Flash-Next	Intel Arc B570 10GB	9.0	150W	6	30.8
109	GLM 5.3	2× DGX Station GB300	81.5	1400W	5.8	3.41
110	Qwen3.8-27B TurboFCFusion	RTX 5060 Ti 16GB	10	180W	5.6	27.78
111	REAP50 GLM 5.3 Flash	Radeon RX 9070 XT	10	304W	3.3	27.78