@sylphai/adal-cli-linux-x64 0.1.3-beta.32 → 0.1.3-beta.34
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
|
Binary file
|
|
@@ -0,0 +1,368 @@
|
|
|
1
|
+
{
|
|
2
|
+
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
+
"version": "1.0.0",
|
|
4
|
+
"last_updated": "2025-10-13",
|
|
5
|
+
"description": "Centralized API model pricing configuration for deep_research",
|
|
6
|
+
"sync_note": "⚠️ IMPORTANT: This file must be kept in sync with backend/src/adal_backend/services/model_pricing.json. If you update pricing here, update it there too!",
|
|
7
|
+
"sources": {
|
|
8
|
+
"anthropic": "https://www.claude.com/pricing",
|
|
9
|
+
"openai": "https://openai.com/api/pricing/",
|
|
10
|
+
"google": "https://ai.google.dev/pricing",
|
|
11
|
+
"deepseek": "https://www.deepseek.com/pricing",
|
|
12
|
+
"prompt_caching_anthropic": "https://www.anthropic.com/news/prompt-caching"
|
|
13
|
+
},
|
|
14
|
+
"notes": {
|
|
15
|
+
"pricing_unit": "USD per million tokens (MTok)",
|
|
16
|
+
"local_models": "All Ollama models are FREE (no API costs)",
|
|
17
|
+
"cached_discount": "Anthropic: 90% off cached input, OpenAI GPT-5: 90% off, GPT-4o: 50% off",
|
|
18
|
+
"reasoning_tokens": "GPT-5 reasoning tokens billed at input token rate",
|
|
19
|
+
"minimum_cacheable": "1,024 tokens for Anthropic and OpenAI",
|
|
20
|
+
"cache_ttl": "5 minutes for Anthropic (OpenAI not publicly documented)",
|
|
21
|
+
"openai_models": "Only GPT-5 and GPT-4o are officially supported. Other models have been removed."
|
|
22
|
+
},
|
|
23
|
+
"providers": {
|
|
24
|
+
"anthropic": {
|
|
25
|
+
"models": {
|
|
26
|
+
"claude-opus-4-1": {
|
|
27
|
+
"display_name": "Claude Opus 4.1",
|
|
28
|
+
"input_price_per_mtok": 15.0,
|
|
29
|
+
"cached_input_price_per_mtok": 1.5,
|
|
30
|
+
"cache_write_price_per_mtok": 18.75,
|
|
31
|
+
"output_price_per_mtok": 75.0,
|
|
32
|
+
"cache_discount_percent": 90,
|
|
33
|
+
"cache_write_markup_percent": 25,
|
|
34
|
+
"supports_caching": true,
|
|
35
|
+
"supports_thinking": true,
|
|
36
|
+
"model_type": "flagship",
|
|
37
|
+
"notes": "Latest flagship, thinking model"
|
|
38
|
+
},
|
|
39
|
+
"claude-opus-4": {
|
|
40
|
+
"display_name": "Claude Opus 4",
|
|
41
|
+
"input_price_per_mtok": 15.0,
|
|
42
|
+
"cached_input_price_per_mtok": 1.5,
|
|
43
|
+
"cache_write_price_per_mtok": 18.75,
|
|
44
|
+
"output_price_per_mtok": 75.0,
|
|
45
|
+
"cache_discount_percent": 90,
|
|
46
|
+
"cache_write_markup_percent": 25,
|
|
47
|
+
"supports_caching": true,
|
|
48
|
+
"supports_thinking": true,
|
|
49
|
+
"model_type": "flagship",
|
|
50
|
+
"notes": "Thinking model (extended reasoning)"
|
|
51
|
+
},
|
|
52
|
+
"claude-sonnet-4-5": {
|
|
53
|
+
"display_name": "Claude Sonnet 4.5",
|
|
54
|
+
"input_price_per_mtok": 3.0,
|
|
55
|
+
"cached_input_price_per_mtok": 0.3,
|
|
56
|
+
"cache_write_price_per_mtok": 3.75,
|
|
57
|
+
"output_price_per_mtok": 15.0,
|
|
58
|
+
"cache_discount_percent": 90,
|
|
59
|
+
"cache_write_markup_percent": 25,
|
|
60
|
+
"supports_caching": true,
|
|
61
|
+
"supports_thinking": true,
|
|
62
|
+
"model_type": "mid-tier",
|
|
63
|
+
"notes": "State-of-the-art, thinking model, best value"
|
|
64
|
+
},
|
|
65
|
+
"claude-sonnet-4-5-1m": {
|
|
66
|
+
"display_name": "Claude Sonnet 4.5 (1M)",
|
|
67
|
+
"input_price_per_mtok": 6.0,
|
|
68
|
+
"cached_input_price_per_mtok": 0.6,
|
|
69
|
+
"cache_write_price_per_mtok": 7.5,
|
|
70
|
+
"output_price_per_mtok": 22.5,
|
|
71
|
+
"cache_discount_percent": 90,
|
|
72
|
+
"cache_write_markup_percent": 25,
|
|
73
|
+
"supports_caching": true,
|
|
74
|
+
"supports_thinking": true,
|
|
75
|
+
"model_type": "extended-context",
|
|
76
|
+
"notes": "1M context window variant, 2x base pricing"
|
|
77
|
+
},
|
|
78
|
+
"claude-sonnet-4": {
|
|
79
|
+
"display_name": "Claude Sonnet 4",
|
|
80
|
+
"input_price_per_mtok": 3.0,
|
|
81
|
+
"cached_input_price_per_mtok": 0.3,
|
|
82
|
+
"cache_write_price_per_mtok": 3.75,
|
|
83
|
+
"output_price_per_mtok": 15.0,
|
|
84
|
+
"cache_discount_percent": 90,
|
|
85
|
+
"cache_write_markup_percent": 25,
|
|
86
|
+
"supports_caching": true,
|
|
87
|
+
"supports_thinking": true,
|
|
88
|
+
"model_type": "mid-tier",
|
|
89
|
+
"notes": "Thinking model"
|
|
90
|
+
},
|
|
91
|
+
"claude-3-7-sonnet": {
|
|
92
|
+
"display_name": "Claude 3.7 Sonnet",
|
|
93
|
+
"input_price_per_mtok": 3.0,
|
|
94
|
+
"cached_input_price_per_mtok": 0.3,
|
|
95
|
+
"cache_write_price_per_mtok": 3.75,
|
|
96
|
+
"output_price_per_mtok": 15.0,
|
|
97
|
+
"cache_discount_percent": 90,
|
|
98
|
+
"cache_write_markup_percent": 25,
|
|
99
|
+
"supports_caching": true,
|
|
100
|
+
"supports_thinking": true,
|
|
101
|
+
"model_type": "mid-tier",
|
|
102
|
+
"notes": "Thinking model"
|
|
103
|
+
},
|
|
104
|
+
"claude-3-5-sonnet": {
|
|
105
|
+
"display_name": "Claude 3.5 Sonnet",
|
|
106
|
+
"input_price_per_mtok": 3.0,
|
|
107
|
+
"cached_input_price_per_mtok": 0.3,
|
|
108
|
+
"cache_write_price_per_mtok": 3.75,
|
|
109
|
+
"output_price_per_mtok": 15.0,
|
|
110
|
+
"cache_discount_percent": 90,
|
|
111
|
+
"cache_write_markup_percent": 25,
|
|
112
|
+
"supports_caching": true,
|
|
113
|
+
"supports_thinking": false,
|
|
114
|
+
"model_type": "mid-tier",
|
|
115
|
+
"notes": "Standard model (no thinking)"
|
|
116
|
+
},
|
|
117
|
+
"claude-haiku-4-5": {
|
|
118
|
+
"display_name": "Claude Haiku 4.5",
|
|
119
|
+
"input_price_per_mtok": 1.0,
|
|
120
|
+
"cached_input_price_per_mtok": 0.1,
|
|
121
|
+
"cache_write_price_per_mtok": 1.25,
|
|
122
|
+
"output_price_per_mtok": 5.0,
|
|
123
|
+
"cache_discount_percent": 90,
|
|
124
|
+
"cache_write_markup_percent": 25,
|
|
125
|
+
"supports_caching": true,
|
|
126
|
+
"supports_thinking": true,
|
|
127
|
+
"model_type": "budget",
|
|
128
|
+
"notes": "Claude Haiku 4.5 released Oct 2025, $1/$5 per MTok with prompt caching support"
|
|
129
|
+
},
|
|
130
|
+
"claude-3-haiku": {
|
|
131
|
+
"display_name": "Claude 3 Haiku",
|
|
132
|
+
"input_price_per_mtok": 0.8,
|
|
133
|
+
"cached_input_price_per_mtok": 0.08,
|
|
134
|
+
"cache_write_price_per_mtok": 1.0,
|
|
135
|
+
"output_price_per_mtok": 4.0,
|
|
136
|
+
"cache_discount_percent": 90,
|
|
137
|
+
"cache_write_markup_percent": 25,
|
|
138
|
+
"supports_caching": true,
|
|
139
|
+
"supports_thinking": false,
|
|
140
|
+
"model_type": "budget",
|
|
141
|
+
"notes": "Fastest, lightweight model"
|
|
142
|
+
}
|
|
143
|
+
},
|
|
144
|
+
"cache_write_markup": 0.25,
|
|
145
|
+
"cache_ttl_minutes": 5,
|
|
146
|
+
"minimum_cacheable_tokens": 1024
|
|
147
|
+
},
|
|
148
|
+
"openai": {
|
|
149
|
+
"models": {
|
|
150
|
+
"gpt-5": {
|
|
151
|
+
"display_name": "GPT-5",
|
|
152
|
+
"input_price_per_mtok": 1.25,
|
|
153
|
+
"cached_input_price_per_mtok": 0.125,
|
|
154
|
+
"output_price_per_mtok": 10.0,
|
|
155
|
+
"reasoning_price_per_mtok": 1.25,
|
|
156
|
+
"cache_discount_percent": 90,
|
|
157
|
+
"supports_caching": true,
|
|
158
|
+
"supports_thinking": true,
|
|
159
|
+
"model_type": "flagship",
|
|
160
|
+
"notes": "Thinking model, reasoning effort configurable, cheapest flagship"
|
|
161
|
+
},
|
|
162
|
+
"gpt-5-codex": {
|
|
163
|
+
"display_name": "GPT-5-Codex",
|
|
164
|
+
"input_price_per_mtok": 1.25,
|
|
165
|
+
"cached_input_price_per_mtok": 0.125,
|
|
166
|
+
"output_price_per_mtok": 10.0,
|
|
167
|
+
"reasoning_price_per_mtok": 1.25,
|
|
168
|
+
"cache_discount_percent": 90,
|
|
169
|
+
"supports_caching": true,
|
|
170
|
+
"supports_thinking": true,
|
|
171
|
+
"model_type": "specialized",
|
|
172
|
+
"notes": "Coding-specialized variant of GPT-5, same pricing"
|
|
173
|
+
},
|
|
174
|
+
"gpt-5.1": {
|
|
175
|
+
"display_name": "GPT-5.1",
|
|
176
|
+
"input_price_per_mtok": 1.25,
|
|
177
|
+
"cached_input_price_per_mtok": 0.125,
|
|
178
|
+
"output_price_per_mtok": 10.0,
|
|
179
|
+
"reasoning_price_per_mtok": 1.25,
|
|
180
|
+
"cache_discount_percent": 90,
|
|
181
|
+
"supports_caching": true,
|
|
182
|
+
"supports_thinking": true,
|
|
183
|
+
"model_type": "flagship",
|
|
184
|
+
"notes": "Adaptive reasoning, 196K context, 24h cache TTL, improved over GPT-5"
|
|
185
|
+
},
|
|
186
|
+
"gpt-4.1": {
|
|
187
|
+
"display_name": "GPT-4.1",
|
|
188
|
+
"input_price_per_mtok": 2.0,
|
|
189
|
+
"cached_input_price_per_mtok": 0.5,
|
|
190
|
+
"output_price_per_mtok": 8.0,
|
|
191
|
+
"cache_discount_percent": 75,
|
|
192
|
+
"supports_caching": true,
|
|
193
|
+
"supports_thinking": false,
|
|
194
|
+
"model_type": "flagship",
|
|
195
|
+
"notes": "1M context window, 75% cache discount, optimized for coding"
|
|
196
|
+
},
|
|
197
|
+
"gpt-4o": {
|
|
198
|
+
"display_name": "GPT-4o",
|
|
199
|
+
"input_price_per_mtok": 2.5,
|
|
200
|
+
"cached_input_price_per_mtok": 1.25,
|
|
201
|
+
"output_price_per_mtok": 10.0,
|
|
202
|
+
"cache_discount_percent": 50,
|
|
203
|
+
"supports_caching": true,
|
|
204
|
+
"supports_thinking": false,
|
|
205
|
+
"model_type": "flagship",
|
|
206
|
+
"notes": "Standard model, no extended reasoning"
|
|
207
|
+
}
|
|
208
|
+
},
|
|
209
|
+
"minimum_cacheable_tokens": 1024
|
|
210
|
+
},
|
|
211
|
+
"google": {
|
|
212
|
+
"models": {
|
|
213
|
+
"gemini-3-pro-preview": {
|
|
214
|
+
"display_name": "Gemini 3 Pro (Preview)",
|
|
215
|
+
"input_price_per_mtok": 2.0,
|
|
216
|
+
"cached_input_price_per_mtok": 0.5,
|
|
217
|
+
"output_price_per_mtok": 12.0,
|
|
218
|
+
"cache_discount_percent": 75,
|
|
219
|
+
"supports_caching": true,
|
|
220
|
+
"supports_thinking": true,
|
|
221
|
+
"notes": "Latest flagship model (Nov 2025), 1M context window, state-of-the-art reasoning. Best model in the world for multimodal understanding"
|
|
222
|
+
},
|
|
223
|
+
"gemini-2.5-pro": {
|
|
224
|
+
"display_name": "Gemini 2.5 Pro",
|
|
225
|
+
"input_price_per_mtok": 1.25,
|
|
226
|
+
"cached_input_price_per_mtok": 0.3125,
|
|
227
|
+
"output_price_per_mtok": 10.0,
|
|
228
|
+
"cache_discount_percent": 75,
|
|
229
|
+
"supports_caching": true,
|
|
230
|
+
"supports_thinking": true,
|
|
231
|
+
"notes": "Context <200K: $1.25 input, $10 output; >200K: $2.50 input, $15 output. Using base rate for simplicity. Cached rate estimated based on 75% discount"
|
|
232
|
+
},
|
|
233
|
+
"gemini-2.5-flash": {
|
|
234
|
+
"display_name": "Gemini 2.5 Flash",
|
|
235
|
+
"input_price_per_mtok": 0.30,
|
|
236
|
+
"cached_input_price_per_mtok": 0.075,
|
|
237
|
+
"output_price_per_mtok": 2.50,
|
|
238
|
+
"cache_discount_percent": 75,
|
|
239
|
+
"batch_discount_percent": 50,
|
|
240
|
+
"supports_caching": true,
|
|
241
|
+
"supports_thinking": true,
|
|
242
|
+
"notes": "Budget-friendly, 1M context, thinking mode, batch: $0.05 input/$0.20 output; cached rate estimated based on 75% discount"
|
|
243
|
+
},
|
|
244
|
+
"gemini-2.0-flash-exp": {
|
|
245
|
+
"display_name": "Gemini 2.0 Flash (Experimental)",
|
|
246
|
+
"input_price_per_mtok": 0.0,
|
|
247
|
+
"output_price_per_mtok": 0.0,
|
|
248
|
+
"cached_input_price_per_mtok": 0.0,
|
|
249
|
+
"supports_caching": false,
|
|
250
|
+
"supports_thinking": false,
|
|
251
|
+
"model_type": "experimental",
|
|
252
|
+
"notes": "Free experimental model"
|
|
253
|
+
},
|
|
254
|
+
"gemini-1.5-pro": {
|
|
255
|
+
"display_name": "Gemini 1.5 Pro",
|
|
256
|
+
"input_price_per_mtok": 1.25,
|
|
257
|
+
"cached_input_price_per_mtok": 0.3125,
|
|
258
|
+
"output_price_per_mtok": 5.0,
|
|
259
|
+
"cache_discount_percent": 75,
|
|
260
|
+
"supports_caching": true,
|
|
261
|
+
"supports_thinking": false,
|
|
262
|
+
"model_type": "flagship"
|
|
263
|
+
},
|
|
264
|
+
"gemini-1.5-flash": {
|
|
265
|
+
"display_name": "Gemini 1.5 Flash",
|
|
266
|
+
"input_price_per_mtok": 0.075,
|
|
267
|
+
"cached_input_price_per_mtok": 0.01875,
|
|
268
|
+
"output_price_per_mtok": 0.3,
|
|
269
|
+
"cache_discount_percent": 75,
|
|
270
|
+
"supports_caching": true,
|
|
271
|
+
"supports_thinking": false,
|
|
272
|
+
"model_type": "mid-tier"
|
|
273
|
+
},
|
|
274
|
+
"gemini-1.5-flash-8b": {
|
|
275
|
+
"display_name": "Gemini 1.5 Flash 8B",
|
|
276
|
+
"input_price_per_mtok": 0.0375,
|
|
277
|
+
"cached_input_price_per_mtok": 0.009375,
|
|
278
|
+
"output_price_per_mtok": 0.15,
|
|
279
|
+
"cache_discount_percent": 75,
|
|
280
|
+
"supports_caching": true,
|
|
281
|
+
"supports_thinking": false,
|
|
282
|
+
"model_type": "budget"
|
|
283
|
+
}
|
|
284
|
+
}
|
|
285
|
+
},
|
|
286
|
+
"deepseek": {
|
|
287
|
+
"models": {
|
|
288
|
+
"deepseek-chat": {
|
|
289
|
+
"display_name": "DeepSeek Chat",
|
|
290
|
+
"input_price_per_mtok": 0.27,
|
|
291
|
+
"cached_input_price_per_mtok": 0.014,
|
|
292
|
+
"output_price_per_mtok": 1.1,
|
|
293
|
+
"cache_discount_percent": 95,
|
|
294
|
+
"supports_caching": true,
|
|
295
|
+
"supports_thinking": false,
|
|
296
|
+
"model_type": "mid-tier"
|
|
297
|
+
},
|
|
298
|
+
"deepseek-reasoner": {
|
|
299
|
+
"display_name": "DeepSeek Reasoner",
|
|
300
|
+
"input_price_per_mtok": 0.55,
|
|
301
|
+
"cached_input_price_per_mtok": 0.014,
|
|
302
|
+
"output_price_per_mtok": 2.19,
|
|
303
|
+
"cache_discount_percent": 97,
|
|
304
|
+
"supports_caching": true,
|
|
305
|
+
"supports_thinking": true,
|
|
306
|
+
"model_type": "thinking"
|
|
307
|
+
}
|
|
308
|
+
}
|
|
309
|
+
},
|
|
310
|
+
"ollama": {
|
|
311
|
+
"models": {
|
|
312
|
+
"gpt-oss-20b": {
|
|
313
|
+
"display_name": "GPT-OSS 20B",
|
|
314
|
+
"input_price_per_mtok": 0.0,
|
|
315
|
+
"output_price_per_mtok": 0.0,
|
|
316
|
+
"supports_caching": false,
|
|
317
|
+
"supports_thinking": false,
|
|
318
|
+
"model_type": "local",
|
|
319
|
+
"notes": "FREE - Local model, no API costs"
|
|
320
|
+
},
|
|
321
|
+
"qwen3-coder-30b": {
|
|
322
|
+
"display_name": "Qwen3 Coder 30B",
|
|
323
|
+
"input_price_per_mtok": 0.0,
|
|
324
|
+
"output_price_per_mtok": 0.0,
|
|
325
|
+
"supports_caching": false,
|
|
326
|
+
"supports_thinking": false,
|
|
327
|
+
"model_type": "local",
|
|
328
|
+
"notes": "FREE - Local model, no API costs"
|
|
329
|
+
}
|
|
330
|
+
},
|
|
331
|
+
"notes": "All Ollama models are FREE. Skip credit deduction, only save usage with cost=$0, is_local=true"
|
|
332
|
+
}
|
|
333
|
+
},
|
|
334
|
+
"cost_calculation_examples": {
|
|
335
|
+
"claude_sonnet_4_5_with_caching": {
|
|
336
|
+
"description": "100K context (cached) + 2K new input + 5K output",
|
|
337
|
+
"first_call_cache_miss": {
|
|
338
|
+
"input_tokens": 102000,
|
|
339
|
+
"input_cost": 0.306,
|
|
340
|
+
"cache_write_tokens": 100000,
|
|
341
|
+
"cache_write_cost": 0.075,
|
|
342
|
+
"output_tokens": 5000,
|
|
343
|
+
"output_cost": 0.075,
|
|
344
|
+
"total_cost": 0.456
|
|
345
|
+
},
|
|
346
|
+
"subsequent_call_cache_hit": {
|
|
347
|
+
"cached_tokens": 100000,
|
|
348
|
+
"cached_cost": 0.03,
|
|
349
|
+
"new_input_tokens": 2000,
|
|
350
|
+
"new_input_cost": 0.006,
|
|
351
|
+
"output_tokens": 5000,
|
|
352
|
+
"output_cost": 0.075,
|
|
353
|
+
"total_cost": 0.111,
|
|
354
|
+
"savings_percent": 76
|
|
355
|
+
}
|
|
356
|
+
},
|
|
357
|
+
"gpt5_with_reasoning": {
|
|
358
|
+
"description": "50K input + 10K reasoning tokens (medium effort) + 3K output",
|
|
359
|
+
"input_tokens": 50000,
|
|
360
|
+
"input_cost": 0.0625,
|
|
361
|
+
"reasoning_tokens": 10000,
|
|
362
|
+
"reasoning_cost": 0.0125,
|
|
363
|
+
"output_tokens": 3000,
|
|
364
|
+
"output_cost": 0.03,
|
|
365
|
+
"total_cost": 0.105
|
|
366
|
+
}
|
|
367
|
+
}
|
|
368
|
+
}
|
|
Binary file
|