-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathmodels.example.json
More file actions
218 lines (218 loc) · 6.13 KB
/
Copy pathmodels.example.json
File metadata and controls
218 lines (218 loc) · 6.13 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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
{
"models": {
"openai/gpt-5.3-codex": {
"provider": "openai",
"name": "GPT-5.3 Codex",
"context_window": 400000,
"cost": {
"input_per_1m": 1.75,
"cache_read_per_1m": 0.175,
"cache_write_per_1m": 1.75,
"output_per_1m": 14
},
"strengths": [
"planning and orchestration",
"architecture decisions",
"long-context investigation",
"code generation and refactoring",
"analytical reasoning"
]
},
"github-copilot/gpt-5.3-codex": {
"provider": "github-copilot",
"name": "GPT-5.3 Codex (Copilot)",
"context_window": 400000,
"cost": {
"input_per_1m": 1.75,
"cache_read_per_1m": 0.175,
"cache_write_per_1m": 1.75,
"output_per_1m": 14
},
"strengths": [
"planning",
"code generation"
],
"notes": "Prefer openai/gpt-5.3-codex for the larger context window."
},
"github-copilot/gpt-5.4-mini": {
"provider": "github-copilot",
"name": "GPT-5.4 Mini (Copilot)",
"context_window": 400000,
"cost": {
"input_per_1m": 0.75,
"cache_read_per_1m": 0.075,
"cache_write_per_1m": 0.75,
"output_per_1m": 4.5
},
"strengths": [
"lightweight exploration",
"compact summarization",
"low-latency tool orchestration"
]
},
"github-copilot/gpt-4o": {
"provider": "github-copilot",
"name": "GPT-4o (Copilot)",
"context_window": 128000,
"cost": {
"input_per_1m": "free",
"output_per_1m": "free"
},
"strengths": [
"fast structured output",
"code review",
"read-only exploration",
"lightweight investigation"
]
},
"github-copilot/gpt-5-mini": {
"provider": "github-copilot",
"name": "GPT-5 Mini (Copilot)",
"context_window": 192000,
"cost": {
"input_per_1m": "free",
"output_per_1m": "free"
},
"strengths": [
"lightweight housekeeping",
"structured formatting",
"minimal tool use"
]
},
"github-copilot/claude-sonnet-4.6": {
"provider": "github-copilot",
"name": "Claude Sonnet 4.6 (Copilot)",
"context_window": 160000,
"cost": {
"input_per_1m": 3,
"cache_read_per_1m": 0.3,
"cache_write_per_1m": 3.75,
"cache_write_5m_per_1m": 3.75,
"cache_write_1h_per_1m": 6,
"output_per_1m": 15
},
"strengths": [
"targeted code implementation",
"refactoring",
"infrastructure-as-code",
"long-form writing and documentation",
"sustained multi-step task execution"
]
},
"github-copilot/claude-opus-4.7": {
"provider": "github-copilot",
"name": "Claude Opus 4.7 (Copilot)",
"context_window": 192000,
"cost": {
"input_per_1m": 5,
"cache_read_per_1m": 0.5,
"cache_write_per_1m": 6.25,
"cache_write_5m_per_1m": 6.25,
"cache_write_1h_per_1m": 10,
"output_per_1m": 25
},
"strengths": [
"deep planning and reasoning",
"complex requirements elicitation",
"nuanced long-form writing"
],
"notes": "Expensive (~15x Sonnet). Reserved for A/B testing and high-stakes planning tasks where reasoning quality is demonstrably better."
},
"github-copilot/claude-haiku-4.5": {
"provider": "github-copilot",
"name": "Claude Haiku 4.5 (Copilot)",
"context_window": 160000,
"cost": {
"input_per_1m": 1,
"cache_read_per_1m": 0.1,
"cache_write_per_1m": 1.25,
"cache_write_5m_per_1m": 1.25,
"cache_write_1h_per_1m": 2,
"output_per_1m": 5
},
"strengths": [
"fast, lightweight tasks",
"structured extraction",
"low-latency tool use"
],
"notes": "Good alternative to gpt-5-mini for structured output tasks."
},
"github-copilot/gemini-3.1-pro-preview": {
"provider": "github-copilot",
"name": "Gemini 3.1 Pro Preview (Copilot)",
"context_window": 173000,
"cost": {
"input_per_1m": 1.02,
"output_per_1m": 12.03
},
"strengths": [
"strong reasoning",
"very long context",
"search and synthesis"
],
"notes": "Preview model — A/B candidate for technical-lead or systems-architect."
},
"ollama/qwen3-coder": {
"provider": "ollama",
"name": "Qwen3 Coder 30B (local)",
"context_window": 32000,
"cost": {
"input_per_1m": 0,
"output_per_1m": 0
},
"strengths": [
"zero cost",
"offline / air-gapped use",
"code generation"
],
"notes": "Requires local Ollama instance at http://localhost:11434."
},
"openai/gpt-5.5": {
"provider": "openai",
"name": "gpt-5.5",
"context_window": 500000,
"cost": {
"input_per_1m": 5,
"cache_read_per_1m": 0.5,
"cache_write_per_1m": 5,
"output_per_1m": 30
},
"strengths": [
"agentic workloads"
],
"notes": ""
}
},
"agent_assignments": {
"technical-lead": "openai/gpt-5.5",
"tpm": "openai/gpt-5.5",
"systems-architect": "openai/gpt-5.5",
"security-engineer": "github-copilot/claude-sonnet-4.6",
"backend-engineer": "github-copilot/claude-sonnet-4.6",
"devops-engineer": "github-copilot/claude-sonnet-4.6",
"qa": "github-copilot/claude-sonnet-4.6",
"task-resolver": "openai/gpt-5.3-codex",
"task-refiner": "openai/gpt-5.3-codex",
"issue-reporter": "openai/gpt-5.3-codex",
"plan": "openai/gpt-5.5",
"build": "openai/gpt-5.3-codex",
"explore": "openai/gpt-5.3-codex",
"title": "github-copilot/gpt-4o",
"summary": "github-copilot/gpt-5.4-mini",
"compaction": "github-copilot/gpt-5.4-mini"
},
"ab_test_candidates": {
"technical-lead": [
"github-copilot/gemini-3.1-pro-preview"
],
"tpm": [
"github-copilot/claude-opus-4.7"
],
"qa": [
"openai/gpt-5.3-codex"
],
"issue-reporter": [
"github-copilot/claude-haiku-4.5"
]
}
}