Repository navigation
Expand file tree
/
Copy pathcost_analysis_output.txt
More file actions
150 lines (131 loc) · 8.39 KB
/
Copy pathcost_analysis_output.txt
File metadata and controls
150 lines (131 loc) · 8.39 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
STG 成本分析工具
================================================================================
[场景 1] 实测数据（31,077 tokens 输入）
--------------------------------------------------------------------------------
================================================================================
成本分析报告
================================================================================
目标模型: Claude Opus 4.6
原始输入 tokens: 31,077
输出 tokens: 1,000
压缩触发: 是
压缩后 tokens: 12,120
压缩器消耗: 14,916
--------------------------------------------------------------------------------
场景 输入成本 输出成本 总成本
--------------------------------------------------------------------------------
不使用 STG $0.4662 $0.0750 $0.5412
压缩器成本 $0.0045 - $0.0045
压缩后输入 $0.1818 - $0.1818
输出成本 - $0.0750 $0.0750
使用 STG 总计 - - $0.2613
--------------------------------------------------------------------------------
[成本节省]
- 节省金额: $0.2799 (51.7%)
- 节省 tokens: 18,957 (61.0%)
- ROI: 62.55x (每花费 $1 压缩，节省 $62.55)
================================================================================
================================================================================
成本分析报告
================================================================================
目标模型: Claude Sonnet 4
原始输入 tokens: 31,077
输出 tokens: 1,000
压缩触发: 是
压缩后 tokens: 12,120
压缩器消耗: 14,916
--------------------------------------------------------------------------------
场景 输入成本 输出成本 总成本
--------------------------------------------------------------------------------
不使用 STG $0.0932 $0.0150 $0.1082
压缩器成本 $0.0045 - $0.0045
压缩后输入 $0.0364 - $0.0364
输出成本 - $0.0150 $0.0150
使用 STG 总计 - - $0.0558
--------------------------------------------------------------------------------
[成本节省]
- 节省金额: $0.0524 (48.4%)
- 节省 tokens: 18,957 (61.0%)
- ROI: 11.71x (每花费 $1 压缩，节省 $11.71)
================================================================================
================================================================================
成本分析报告
================================================================================
目标模型: GPT-4
原始输入 tokens: 31,077
输出 tokens: 1,000
压缩触发: 是
压缩后 tokens: 12,120
压缩器消耗: 14,916
--------------------------------------------------------------------------------
场景 输入成本 输出成本 总成本
--------------------------------------------------------------------------------
不使用 STG $0.3108 $0.0300 $0.3408
压缩器成本 $0.0045 - $0.0045
压缩后输入 $0.1212 - $0.1212
输出成本 - $0.0300 $0.0300
使用 STG 总计 - - $0.1557
--------------------------------------------------------------------------------
[成本节省]
- 节省金额: $0.1851 (54.3%)
- 节省 tokens: 18,957 (61.0%)
- ROI: 41.36x (每花费 $1 压缩，节省 $41.36)
================================================================================
================================================================================
成本分析报告
================================================================================
目标模型: Qwen 3.5 27B
原始输入 tokens: 31,077
输出 tokens: 1,000
压缩触发: 是
压缩后 tokens: 12,120
压缩器消耗: 14,916
--------------------------------------------------------------------------------
场景 输入成本 输出成本 总成本
--------------------------------------------------------------------------------
不使用 STG $0.0093 $0.0024 $0.0117
压缩器成本 $0.0045 - $0.0045
压缩后输入 $0.0036 - $0.0036
输出成本 - $0.0024 $0.0024
使用 STG 总计 - - $0.0105
--------------------------------------------------------------------------------
[成本节省]
- 节省金额: $0.0012 (10.3%)
- 节省 tokens: 18,957 (61.0%)
- ROI: 0.27x (每花费 $1 压缩，节省 $0.27)
================================================================================
[场景 2] Token 数量 vs 成本节省曲线
--------------------------------------------------------------------------------
[OK] 曲线数据已保存到 cost_curve_data.json
========================================================================================================================
Token 数量 模型 不使用 STG 使用 STG 节省金额 节省比例 ROI
========================================================================================================================
2,000 opus-4.6 $0.1050 $0.1050 $0.0000 0.0 % 0.00 x
4,096 opus-4.6 $0.1364 $0.0995 $0.0369 27.0 % 62.56 x
8,000 opus-4.6 $0.1950 $0.1230 $0.0720 36.9 % 62.54 x
16,000 opus-4.6 $0.3150 $0.1709 $0.1441 45.7 % 62.54 x
32,000 opus-4.6 $0.5550 $0.2668 $0.2882 51.9 % 62.54 x
64,000 opus-4.6 $1.0350 $0.4586 $0.5764 55.7 % 62.54 x
128,000 opus-4.6 $1.9950 $0.8422 $1.1528 57.8 % 62.54 x
2,000 sonnet-4 $0.0210 $0.0210 $0.0000 0.0 % 0.00 x
4,096 sonnet-4 $0.0273 $0.0204 $0.0069 25.3 % 11.71 x
8,000 sonnet-4 $0.0390 $0.0255 $0.0135 34.6 % 11.71 x
16,000 sonnet-4 $0.0630 $0.0360 $0.0270 42.8 % 11.71 x
32,000 sonnet-4 $0.1110 $0.0570 $0.0540 48.6 % 11.71 x
64,000 sonnet-4 $0.2070 $0.0991 $0.1079 52.1 % 11.71 x
128,000 sonnet-4 $0.3990 $0.1832 $0.2158 54.1 % 11.71 x
2,000 gpt-4 $0.0500 $0.0500 $0.0000 0.0 % 0.00 x
4,096 gpt-4 $0.0710 $0.0466 $0.0244 34.4 % 41.37 x
8,000 gpt-4 $0.1100 $0.0624 $0.0476 43.3 % 41.36 x
16,000 gpt-4 $0.1900 $0.0947 $0.0953 50.2 % 41.36 x
32,000 gpt-4 $0.3500 $0.1594 $0.1906 54.5 % 41.36 x
64,000 gpt-4 $0.6700 $0.2888 $0.3812 56.9 % 41.36 x
128,000 gpt-4 $1.3100 $0.5476 $0.7624 58.2 % 41.36 x
2,000 qwen3.5-27b $0.0030 $0.0030 $0.0000 0.0 % 0.00 x
4,096 qwen3.5-27b $0.0036 $0.0035 $0.0002 4.4 % 0.27 x
8,000 qwen3.5-27b $0.0048 $0.0045 $0.0003 6.5 % 0.27 x
16,000 qwen3.5-27b $0.0072 $0.0066 $0.0006 8.7 % 0.27 x
32,000 qwen3.5-27b $0.0120 $0.0108 $0.0012 10.4 % 0.27 x
64,000 qwen3.5-27b $0.0216 $0.0191 $0.0025 11.6 % 0.27 x
128,000 qwen3.5-27b $0.0408 $0.0358 $0.0050 12.2 % 0.27 x
========================================================================================================================