forked from octos-org/octos
-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathconfig.example.json
More file actions
143 lines (143 loc) · 4.79 KB
/
Copy pathconfig.example.json
File metadata and controls
143 lines (143 loc) · 4.79 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
{
"provider": "anthropic",
"model": "claude-sonnet-4-20250514",
"memory": {
"max_inject_tokens": 2500,
"refresh": {
"enabled": true,
"extract_model": null,
"min_idle_minutes": 30,
"max_session_age_days": 10,
"max_sessions_per_pass": 2,
"max_extractions_per_day": 20,
"max_daily_tokens": 200000,
"consolidate_interval_minutes": 30,
"max_extract_input_tokens": 24000,
"consolidate_model": null,
"max_consolidations_per_day": 12,
"debounce_seconds": 90,
"unused_days": 30,
"max_memory_file_tokens": 8000,
"pending_confirm_days": 7
}
},
"sub_providers": [
{
"key": "cheap",
"provider": "openai",
"model": "gpt-4o-mini",
"api_key_env": "OPENAI_API_KEY",
"description": "Fast and cheap. Use for simple search, file ops, and lightweight tasks.",
"default_context_window": 16000
},
{
"key": "strong",
"provider": "openai",
"model": "gpt-4o",
"api_key_env": "OPENAI_API_KEY",
"description": "Strong general-purpose model. Use for code generation and multi-step reasoning."
},
{
"key": "deepseek",
"provider": "deepseek",
"model": "deepseek-v4-pro",
"api_key_env": "DEEPSEEK_API_KEY",
"description": "DeepSeek V4 Pro. Strong general-purpose model for code generation and technical tasks.",
"default_context_window": 1048576
},
{
"key": "deepseek-flash",
"provider": "deepseek",
"model": "deepseek-v4-flash",
"api_key_env": "DEEPSEEK_API_KEY",
"description": "DeepSeek V4 Flash. Fast, low-cost variant for high-throughput coding and chat.",
"default_context_window": 1048576
},
{
"key": "qwen",
"provider": "dashscope",
"model": "qwen-max",
"api_key_env": "DASHSCOPE_API_KEY",
"description": "Alibaba Qwen. Strong multilingual and Chinese-language tasks."
},
{
"key": "qwen-fast",
"provider": "dashscope",
"model": "qwen-turbo",
"api_key_env": "DASHSCOPE_API_KEY",
"description": "Fast Qwen model. Use for quick tasks and large document processing.",
"default_context_window": 16000
},
{
"key": "nvidia-llama",
"provider": "nvidia",
"model": "meta/llama-3.3-70b-instruct",
"api_key_env": "NVIDIA_API_KEY",
"description": "Llama 3.3 70B via NVIDIA NIM. Strong open-source model for general tasks."
},
{
"key": "nvidia-deepseek-v4",
"provider": "nvidia",
"model": "deepseek-ai/deepseek-v4-pro",
"api_key_env": "NVIDIA_API_KEY",
"description": "DeepSeek V4 Pro via NVIDIA NIM. State-of-the-art reasoning for math and coding."
},
{
"key": "nvidia-qwen-coder",
"provider": "nvidia",
"model": "qwen/qwen3-coder-480b-a35b-instruct",
"api_key_env": "NVIDIA_API_KEY",
"description": "Qwen3 Coder 480B via NVIDIA NIM. Top-tier agentic coding model, 256k context."
},
{
"key": "nvidia-mistral",
"provider": "nvidia",
"model": "mistralai/mistral-small-3.1-24b-instruct-2503",
"api_key_env": "NVIDIA_API_KEY",
"description": "Mistral Small 3.1 via NVIDIA NIM. Fast multimodal model, 128k context."
},
{
"key": "kimi",
"provider": "moonshot",
"model": "kimi-k2.5",
"api_key_env": "MOONSHOT_API_KEY",
"description": "Kimi K2.5 by Moonshot. Strong reasoning model with long context, good for complex analysis and agentic tasks."
},
{
"key": "minimax",
"provider": "minimax",
"model": "MiniMax-M3",
"api_key_env": "MINIMAX_API_KEY",
"description": "MiniMax M3. 1M-context multimodal model, strong at math, coding, and agentic tasks.",
"default_context_window": 1048576
},
{
"key": "glm",
"provider": "zhipu",
"model": "glm-5",
"api_key_env": "ZHIPU_API_KEY",
"description": "GLM-5 by Zhipu. 744B MoE reasoning model, strong at agentic tasks and long-horizon planning."
},
{
"key": "qwen-reasoning",
"provider": "dashscope",
"model": "qwq-plus",
"api_key_env": "DASHSCOPE_API_KEY",
"description": "Qwen QwQ-Plus reasoning model. Deep thinking for math, logic, and complex problem-solving."
},
{
"key": "local",
"provider": "ollama",
"model": "qwen3:8b",
"description": "Qwen3 8B running locally via Ollama. Free, private, no API key needed. Good for simple tasks, drafting, and quick iteration.",
"default_context_window": 8000
},
{
"key": "local-14b",
"provider": "ollama",
"model": "qwen3:14b",
"description": "Qwen3 14B running locally via Ollama. Free, private, stronger reasoning than 8B. Use for code review, moderate complexity tasks.",
"default_context_window": 16000
}
]
}