@dirgha/code 1.16.2 → 1.17.1

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.
Files changed (139) hide show
  1. package/dist/cli/config.d.ts +3 -3
  2. package/dist/cli/config.js +34 -14
  3. package/dist/cli/config.js.map +1 -1
  4. package/dist/cli/first-run.d.ts +10 -0
  5. package/dist/cli/first-run.js +86 -0
  6. package/dist/cli/first-run.js.map +1 -0
  7. package/dist/cli/interactive.js +129 -116
  8. package/dist/cli/interactive.js.map +1 -1
  9. package/dist/cli/main.js +24 -4
  10. package/dist/cli/main.js.map +1 -1
  11. package/dist/cli/slash/cost.d.ts +6 -0
  12. package/dist/cli/slash/cost.js +12 -0
  13. package/dist/cli/slash/cost.js.map +1 -0
  14. package/dist/cli/slash/export.d.ts +2 -0
  15. package/dist/cli/slash/export.js +151 -0
  16. package/dist/cli/slash/export.js.map +1 -0
  17. package/dist/cli/slash/fs.d.ts +2 -0
  18. package/dist/cli/slash/fs.js +145 -0
  19. package/dist/cli/slash/fs.js.map +1 -0
  20. package/dist/cli/slash/index.d.ts +29 -25
  21. package/dist/cli/slash/index.js +32 -24
  22. package/dist/cli/slash/index.js.map +1 -1
  23. package/dist/cli/slash/mcp.d.ts +2 -0
  24. package/dist/cli/slash/mcp.js +93 -0
  25. package/dist/cli/slash/mcp.js.map +1 -0
  26. package/dist/cli/slash/models.d.ts +1 -1
  27. package/dist/cli/slash/models.js +21 -20
  28. package/dist/cli/slash/models.js.map +1 -1
  29. package/dist/cli/slash.d.ts +5 -5
  30. package/dist/cli/slash.js +53 -22
  31. package/dist/cli/slash.js.map +1 -1
  32. package/dist/cli/subcommands/models.js +6 -2
  33. package/dist/cli/subcommands/models.js.map +1 -1
  34. package/dist/cli/update-check.d.ts +1 -0
  35. package/dist/cli/update-check.js +77 -0
  36. package/dist/cli/update-check.js.map +1 -0
  37. package/dist/context/default-soul.md +1 -1
  38. package/dist/context/ledger.d.ts +12 -2
  39. package/dist/context/ledger.js +92 -32
  40. package/dist/context/ledger.js.map +1 -1
  41. package/dist/context/memory.d.ts +7 -2
  42. package/dist/context/memory.js +71 -50
  43. package/dist/context/memory.js.map +1 -1
  44. package/dist/context/primer.d.ts +4 -2
  45. package/dist/context/primer.js +18 -10
  46. package/dist/context/primer.js.map +1 -1
  47. package/dist/context/session.d.ts +8 -7
  48. package/dist/context/session.js +40 -23
  49. package/dist/context/session.js.map +1 -1
  50. package/dist/daemon/server.d.ts +4 -4
  51. package/dist/daemon/server.js +66 -36
  52. package/dist/daemon/server.js.map +1 -1
  53. package/dist/fleet/templates.d.ts +18 -0
  54. package/dist/fleet/templates.js +103 -0
  55. package/dist/fleet/templates.js.map +1 -0
  56. package/dist/intelligence/error-classifier.d.ts +2 -2
  57. package/dist/intelligence/error-classifier.js +76 -33
  58. package/dist/intelligence/error-classifier.js.map +1 -1
  59. package/dist/intelligence/prices.js +819 -193
  60. package/dist/intelligence/prices.js.map +1 -1
  61. package/dist/intelligence/smart-router.d.ts +3 -1
  62. package/dist/intelligence/smart-router.js +34 -12
  63. package/dist/intelligence/smart-router.js.map +1 -1
  64. package/dist/kernel/agent-loop.d.ts +11 -0
  65. package/dist/kernel/agent-loop.js +43 -2
  66. package/dist/kernel/agent-loop.js.map +1 -1
  67. package/dist/kernel/types.d.ts +30 -27
  68. package/dist/lsp/client.d.ts +76 -0
  69. package/dist/lsp/client.js +332 -0
  70. package/dist/lsp/client.js.map +1 -0
  71. package/dist/lsp/detector.d.ts +13 -0
  72. package/dist/lsp/detector.js +152 -0
  73. package/dist/lsp/detector.js.map +1 -0
  74. package/dist/lsp/index.d.ts +25 -0
  75. package/dist/lsp/index.js +127 -0
  76. package/dist/lsp/index.js.map +1 -0
  77. package/dist/providers/dispatch.d.ts +1 -1
  78. package/dist/providers/dispatch.js +70 -31
  79. package/dist/providers/dispatch.js.map +1 -1
  80. package/dist/providers/family-fallback.d.ts +12 -0
  81. package/dist/providers/family-fallback.js +73 -0
  82. package/dist/providers/family-fallback.js.map +1 -0
  83. package/dist/providers/health.d.ts +13 -0
  84. package/dist/providers/health.js +71 -0
  85. package/dist/providers/health.js.map +1 -0
  86. package/dist/providers/index.d.ts +15 -12
  87. package/dist/providers/index.js +63 -40
  88. package/dist/providers/index.js.map +1 -1
  89. package/dist/subagents/delegator.d.ts +2 -2
  90. package/dist/subagents/delegator.js +29 -14
  91. package/dist/subagents/delegator.js.map +1 -1
  92. package/dist/subagents/loop-detector.d.ts +30 -0
  93. package/dist/subagents/loop-detector.js +62 -0
  94. package/dist/subagents/loop-detector.js.map +1 -0
  95. package/dist/tools/exec.d.ts +8 -4
  96. package/dist/tools/exec.js +10 -1
  97. package/dist/tools/exec.js.map +1 -1
  98. package/dist/tools/fs-edit.d.ts +1 -1
  99. package/dist/tools/fs-edit.js +30 -17
  100. package/dist/tools/fs-edit.js.map +1 -1
  101. package/dist/tools/index.d.ts +26 -18
  102. package/dist/tools/index.js +31 -18
  103. package/dist/tools/index.js.map +1 -1
  104. package/dist/tools/lsp-definition.d.ts +2 -0
  105. package/dist/tools/lsp-definition.js +48 -0
  106. package/dist/tools/lsp-definition.js.map +1 -0
  107. package/dist/tools/lsp-hover.d.ts +2 -0
  108. package/dist/tools/lsp-hover.js +68 -0
  109. package/dist/tools/lsp-hover.js.map +1 -0
  110. package/dist/tools/lsp-references.d.ts +2 -0
  111. package/dist/tools/lsp-references.js +48 -0
  112. package/dist/tools/lsp-references.js.map +1 -0
  113. package/dist/tools/lsp-symbols.d.ts +2 -0
  114. package/dist/tools/lsp-symbols.js +62 -0
  115. package/dist/tools/lsp-symbols.js.map +1 -0
  116. package/dist/tools/registry.d.ts +3 -2
  117. package/dist/tools/registry.js.map +1 -1
  118. package/dist/tools/shell.d.ts +1 -1
  119. package/dist/tools/shell.js +55 -36
  120. package/dist/tools/shell.js.map +1 -1
  121. package/dist/tui/ink/App.d.ts +2 -0
  122. package/dist/tui/ink/App.js +29 -9
  123. package/dist/tui/ink/App.js.map +1 -1
  124. package/dist/tui/ink/components/ToolBox.d.ts +5 -0
  125. package/dist/tui/ink/components/ToolBox.js +45 -4
  126. package/dist/tui/ink/components/ToolBox.js.map +1 -1
  127. package/dist/tui/ink/components/ToolGroup.d.ts +1 -0
  128. package/dist/tui/ink/components/ToolGroup.js +20 -2
  129. package/dist/tui/ink/components/ToolGroup.js.map +1 -1
  130. package/dist/tui/ink/components/vim-bindings.d.ts +1 -1
  131. package/dist/tui/ink/components/vim-bindings.js +107 -40
  132. package/dist/tui/ink/components/vim-bindings.js.map +1 -1
  133. package/dist/tui/ink/index.d.ts +19 -18
  134. package/dist/tui/ink/index.js +26 -19
  135. package/dist/tui/ink/index.js.map +1 -1
  136. package/dist/tui/ink/use-event-projection.d.ts +2 -0
  137. package/dist/tui/ink/use-event-projection.js +28 -1
  138. package/dist/tui/ink/use-event-projection.js.map +1 -1
  139. package/package.json +1 -1
@@ -10,142 +10,754 @@
10
10
  // charges (their margin is included).
11
11
  export const PRICES = [
12
12
  // Anthropic — native API
13
- { provider: 'anthropic', model: 'claude-opus-4-7', inputPerM: 15, outputPerM: 75, cachedInputPerM: 1.5 },
14
- { provider: 'anthropic', model: 'claude-sonnet-4-6', inputPerM: 3, outputPerM: 15, cachedInputPerM: 0.3 },
15
- { provider: 'anthropic', model: 'claude-haiku-4-5', inputPerM: 0.8, outputPerM: 4, cachedInputPerM: 0.08 },
13
+ {
14
+ provider: "anthropic",
15
+ model: "claude-opus-4-7",
16
+ inputPerM: 15,
17
+ outputPerM: 75,
18
+ cachedInputPerM: 1.5,
19
+ },
20
+ {
21
+ provider: "anthropic",
22
+ model: "claude-sonnet-4-6",
23
+ inputPerM: 3,
24
+ outputPerM: 15,
25
+ cachedInputPerM: 0.3,
26
+ },
27
+ {
28
+ provider: "anthropic",
29
+ model: "claude-haiku-4-5",
30
+ inputPerM: 0.8,
31
+ outputPerM: 4,
32
+ cachedInputPerM: 0.08,
33
+ },
16
34
  // OpenAI — native API
17
- { provider: 'openai', model: 'gpt-5.5-pro', inputPerM: 30, outputPerM: 180 },
18
- { provider: 'openai', model: 'gpt-5.5', inputPerM: 5, outputPerM: 30 },
19
- { provider: 'openai', model: 'gpt-5', inputPerM: 10, outputPerM: 30 },
20
- { provider: 'openai', model: 'gpt-5-mini', inputPerM: 0.5, outputPerM: 2 },
21
- { provider: 'openai', model: 'gpt-4o-mini', inputPerM: 0.15, outputPerM: 0.6 },
22
- { provider: 'openai', model: 'o1', inputPerM: 15, outputPerM: 60 },
35
+ { provider: "openai", model: "gpt-5.5-pro", inputPerM: 30, outputPerM: 180 },
36
+ { provider: "openai", model: "gpt-5.5", inputPerM: 5, outputPerM: 30 },
37
+ { provider: "openai", model: "gpt-5", inputPerM: 10, outputPerM: 30 },
38
+ { provider: "openai", model: "gpt-5-mini", inputPerM: 0.5, outputPerM: 2 },
39
+ {
40
+ provider: "openai",
41
+ model: "gpt-4o-mini",
42
+ inputPerM: 0.15,
43
+ outputPerM: 0.6,
44
+ },
45
+ { provider: "openai", model: "o1", inputPerM: 15, outputPerM: 60 },
23
46
  // Google Gemini — native API (production-stable models)
24
- { provider: 'gemini', model: 'gemini-2.5-pro', inputPerM: 1.25, outputPerM: 5 },
25
- { provider: 'gemini', model: 'gemini-2.5-flash', inputPerM: 0.075, outputPerM: 0.3 },
47
+ {
48
+ provider: "gemini",
49
+ model: "gemini-2.5-pro",
50
+ inputPerM: 1.25,
51
+ outputPerM: 5,
52
+ },
53
+ {
54
+ provider: "gemini",
55
+ model: "gemini-2.5-flash",
56
+ inputPerM: 0.075,
57
+ outputPerM: 0.3,
58
+ },
26
59
  // NVIDIA NIM — verified live on the integrate.api.nvidia.com endpoint
27
- { provider: 'nvidia', model: 'deepseek-ai/deepseek-v4-pro', inputPerM: 0, outputPerM: 0 },
28
- { provider: 'nvidia', model: 'deepseek-ai/deepseek-v4-flash', inputPerM: 0, outputPerM: 0 },
60
+ {
61
+ provider: "nvidia",
62
+ model: "deepseek-ai/deepseek-v4-pro",
63
+ inputPerM: 0,
64
+ outputPerM: 0,
65
+ },
66
+ {
67
+ provider: "nvidia",
68
+ model: "deepseek-ai/deepseek-v4-flash",
69
+ inputPerM: 0,
70
+ outputPerM: 0,
71
+ },
29
72
  // DeepSeek native API (api.deepseek.com) — own quota, no shared 429s.
30
73
  // Cache hits priced at ~10% of base; passed through usage.cached_tokens.
31
- { provider: 'deepseek', model: 'deepseek-chat', inputPerM: 0.27, outputPerM: 1.10, cachedInputPerM: 0.07, contextWindow: 128_000, family: 'deepseek' },
32
- { provider: 'deepseek', model: 'deepseek-v4-flash', inputPerM: 0.07, outputPerM: 0.28, cachedInputPerM: 0.02, contextWindow: 128_000, family: 'deepseek' },
33
- { provider: 'deepseek', model: 'deepseek-v4-pro', inputPerM: 0.43, outputPerM: 0.87, cachedInputPerM: 0.11, contextWindow: 128_000, family: 'deepseek' },
34
- { provider: 'deepseek', model: 'deepseek-reasoner', inputPerM: 0.55, outputPerM: 2.19, cachedInputPerM: 0.14, contextWindow: 128_000, supportsThinking: true, family: 'deepseek' },
35
- { provider: 'deepseek', model: 'deepseek-prover-v2', inputPerM: 0.55, outputPerM: 2.19, cachedInputPerM: 0.14, contextWindow: 128_000, supportsThinking: true, family: 'deepseek' },
36
- { provider: 'nvidia', model: 'moonshotai/kimi-k2-instruct', inputPerM: 0.15, outputPerM: 0.60 },
37
- { provider: 'nvidia', model: 'qwen/qwen3-next-80b-a3b-instruct', inputPerM: 0.08, outputPerM: 0.30 },
38
- { provider: 'nvidia', model: 'meta/llama-3.3-70b-instruct', inputPerM: 0.20, outputPerM: 0.80 },
74
+ {
75
+ provider: "deepseek",
76
+ model: "deepseek-chat",
77
+ inputPerM: 0.27,
78
+ outputPerM: 1.1,
79
+ cachedInputPerM: 0.07,
80
+ contextWindow: 128_000,
81
+ family: "deepseek",
82
+ },
83
+ {
84
+ provider: "deepseek",
85
+ model: "deepseek-v4-flash",
86
+ inputPerM: 0.07,
87
+ outputPerM: 0.28,
88
+ cachedInputPerM: 0.02,
89
+ contextWindow: 128_000,
90
+ family: "deepseek",
91
+ },
92
+ {
93
+ provider: "deepseek",
94
+ model: "deepseek-v4-pro",
95
+ inputPerM: 0.43,
96
+ outputPerM: 0.87,
97
+ cachedInputPerM: 0.11,
98
+ contextWindow: 128_000,
99
+ family: "deepseek",
100
+ },
101
+ {
102
+ provider: "deepseek",
103
+ model: "deepseek-reasoner",
104
+ inputPerM: 0.55,
105
+ outputPerM: 2.19,
106
+ cachedInputPerM: 0.14,
107
+ contextWindow: 128_000,
108
+ supportsThinking: true,
109
+ family: "deepseek",
110
+ },
111
+ {
112
+ provider: "deepseek",
113
+ model: "deepseek-prover-v2",
114
+ inputPerM: 0.55,
115
+ outputPerM: 2.19,
116
+ cachedInputPerM: 0.14,
117
+ contextWindow: 128_000,
118
+ supportsThinking: true,
119
+ family: "deepseek",
120
+ },
121
+ {
122
+ provider: "nvidia",
123
+ model: "moonshotai/kimi-k2-instruct",
124
+ inputPerM: 0.15,
125
+ outputPerM: 0.6,
126
+ },
127
+ {
128
+ provider: "nvidia",
129
+ model: "qwen/qwen3-next-80b-a3b-instruct",
130
+ inputPerM: 0.08,
131
+ outputPerM: 0.3,
132
+ },
133
+ {
134
+ provider: "nvidia",
135
+ model: "meta/llama-3.3-70b-instruct",
136
+ inputPerM: 0.2,
137
+ outputPerM: 0.8,
138
+ },
39
139
  // OpenRouter — premium frontier (priced as OR charges)
40
140
  // 2026-04-28 refresh: `~vendor/...-latest` aliases auto-track the
41
141
  // vendor's newest stable model. Concrete vNs kept for reproducibility.
42
- { provider: 'openrouter', model: '~anthropic/claude-opus-latest', inputPerM: 5, outputPerM: 25 },
43
- { provider: 'openrouter', model: '~anthropic/claude-sonnet-latest', inputPerM: 3, outputPerM: 15 },
44
- { provider: 'openrouter', model: '~anthropic/claude-haiku-latest', inputPerM: 1, outputPerM: 5 },
45
- { provider: 'openrouter', model: 'anthropic/claude-opus-4.7', inputPerM: 5, outputPerM: 25 },
46
- { provider: 'openrouter', model: '~openai/gpt-latest', inputPerM: 5, outputPerM: 30 },
47
- { provider: 'openrouter', model: '~openai/gpt-mini-latest', inputPerM: 0.75, outputPerM: 4.50 },
48
- { provider: 'openrouter', model: 'openai/gpt-5.5-pro', inputPerM: 30, outputPerM: 180 },
49
- { provider: 'openrouter', model: 'openai/gpt-5.5', inputPerM: 5, outputPerM: 30 },
50
- { provider: 'openrouter', model: '~google/gemini-pro-latest', inputPerM: 2, outputPerM: 12 },
51
- { provider: 'openrouter', model: '~google/gemini-flash-latest', inputPerM: 0.5, outputPerM: 3 },
52
- { provider: 'openrouter', model: 'google/gemini-3.1-pro-preview', inputPerM: 2, outputPerM: 12 },
53
- { provider: 'openrouter', model: 'google/gemini-3-flash-preview', inputPerM: 0.5, outputPerM: 3 },
54
- { provider: 'openrouter', model: '~moonshotai/kimi-latest', inputPerM: 0.74, outputPerM: 4.66 },
55
- { provider: 'openrouter', model: 'deepseek/deepseek-v4-pro', inputPerM: 0.43, outputPerM: 0.87 },
56
- { provider: 'openrouter', model: 'deepseek/deepseek-v4-flash', inputPerM: 0.14, outputPerM: 0.28 },
57
- { provider: 'openrouter', model: 'qwen/qwen3.6-max-preview', inputPerM: 1.04, outputPerM: 6.24 },
58
- { provider: 'openrouter', model: 'qwen/qwen3.6-flash', inputPerM: 0.25, outputPerM: 1.50 },
59
- { provider: 'openrouter', model: 'qwen/qwen3.6-35b-a3b', inputPerM: 0.16, outputPerM: 0.97 },
60
- { provider: 'openrouter', model: 'qwen/qwen3.5-plus-20260420', inputPerM: 0.40, outputPerM: 2.40 },
61
- { provider: 'openrouter', model: 'x-ai/grok-4.20-multi-agent', inputPerM: 2, outputPerM: 6 },
62
- { provider: 'openrouter', model: 'x-ai/grok-4.20', inputPerM: 2, outputPerM: 6 },
142
+ {
143
+ provider: "openrouter",
144
+ model: "~anthropic/claude-opus-latest",
145
+ inputPerM: 5,
146
+ outputPerM: 25,
147
+ },
148
+ {
149
+ provider: "openrouter",
150
+ model: "~anthropic/claude-sonnet-latest",
151
+ inputPerM: 3,
152
+ outputPerM: 15,
153
+ },
154
+ {
155
+ provider: "openrouter",
156
+ model: "~anthropic/claude-haiku-latest",
157
+ inputPerM: 1,
158
+ outputPerM: 5,
159
+ },
160
+ {
161
+ provider: "openrouter",
162
+ model: "anthropic/claude-opus-4.7",
163
+ inputPerM: 5,
164
+ outputPerM: 25,
165
+ },
166
+ {
167
+ provider: "openrouter",
168
+ model: "~openai/gpt-latest",
169
+ inputPerM: 5,
170
+ outputPerM: 30,
171
+ },
172
+ {
173
+ provider: "openrouter",
174
+ model: "~openai/gpt-mini-latest",
175
+ inputPerM: 0.75,
176
+ outputPerM: 4.5,
177
+ },
178
+ {
179
+ provider: "openrouter",
180
+ model: "openai/gpt-5.5-pro",
181
+ inputPerM: 30,
182
+ outputPerM: 180,
183
+ },
184
+ {
185
+ provider: "openrouter",
186
+ model: "openai/gpt-5.5",
187
+ inputPerM: 5,
188
+ outputPerM: 30,
189
+ },
190
+ {
191
+ provider: "openrouter",
192
+ model: "~google/gemini-pro-latest",
193
+ inputPerM: 2,
194
+ outputPerM: 12,
195
+ },
196
+ {
197
+ provider: "openrouter",
198
+ model: "~google/gemini-flash-latest",
199
+ inputPerM: 0.5,
200
+ outputPerM: 3,
201
+ },
202
+ {
203
+ provider: "openrouter",
204
+ model: "google/gemini-3.1-pro-preview",
205
+ inputPerM: 2,
206
+ outputPerM: 12,
207
+ },
208
+ {
209
+ provider: "openrouter",
210
+ model: "google/gemini-3-flash-preview",
211
+ inputPerM: 0.5,
212
+ outputPerM: 3,
213
+ },
214
+ {
215
+ provider: "openrouter",
216
+ model: "~moonshotai/kimi-latest",
217
+ inputPerM: 0.74,
218
+ outputPerM: 4.66,
219
+ },
220
+ {
221
+ provider: "openrouter",
222
+ model: "deepseek/deepseek-v4-pro",
223
+ inputPerM: 0.43,
224
+ outputPerM: 0.87,
225
+ },
226
+ {
227
+ provider: "openrouter",
228
+ model: "deepseek/deepseek-v4-flash",
229
+ inputPerM: 0.14,
230
+ outputPerM: 0.28,
231
+ },
232
+ {
233
+ provider: "openrouter",
234
+ model: "qwen/qwen3.6-max-preview",
235
+ inputPerM: 1.04,
236
+ outputPerM: 6.24,
237
+ },
238
+ {
239
+ provider: "openrouter",
240
+ model: "qwen/qwen3.6-flash",
241
+ inputPerM: 0.25,
242
+ outputPerM: 1.5,
243
+ },
244
+ {
245
+ provider: "openrouter",
246
+ model: "qwen/qwen3.6-35b-a3b",
247
+ inputPerM: 0.16,
248
+ outputPerM: 0.97,
249
+ },
250
+ {
251
+ provider: "openrouter",
252
+ model: "qwen/qwen3.5-plus-20260420",
253
+ inputPerM: 0.4,
254
+ outputPerM: 2.4,
255
+ },
256
+ {
257
+ provider: "openrouter",
258
+ model: "x-ai/grok-4.20-multi-agent",
259
+ inputPerM: 2,
260
+ outputPerM: 6,
261
+ },
262
+ {
263
+ provider: "openrouter",
264
+ model: "x-ai/grok-4.20",
265
+ inputPerM: 2,
266
+ outputPerM: 6,
267
+ },
63
268
  // OpenRouter — Moonshot / Kimi
64
- { provider: 'openrouter', model: 'moonshotai/kimi-k2.6', inputPerM: 0.74, outputPerM: 4.66 },
65
- { provider: 'openrouter', model: 'moonshotai/kimi-k2.5', inputPerM: 0.44, outputPerM: 2 },
66
- { provider: 'openrouter', model: 'moonshotai/kimi-k2-thinking', inputPerM: 0.60, outputPerM: 2.50 },
269
+ {
270
+ provider: "openrouter",
271
+ model: "moonshotai/kimi-k2.6",
272
+ inputPerM: 0.74,
273
+ outputPerM: 4.66,
274
+ },
275
+ {
276
+ provider: "openrouter",
277
+ model: "moonshotai/kimi-k2.5",
278
+ inputPerM: 0.44,
279
+ outputPerM: 2,
280
+ },
281
+ {
282
+ provider: "openrouter",
283
+ model: "moonshotai/kimi-k2-thinking",
284
+ inputPerM: 0.6,
285
+ outputPerM: 2.5,
286
+ },
67
287
  // OpenRouter — MiniMax
68
- { provider: 'openrouter', model: 'minimax/minimax-m2.7', inputPerM: 0.30, outputPerM: 1.20 },
69
- { provider: 'openrouter', model: 'minimax/minimax-m2', inputPerM: 0.26, outputPerM: 1 },
288
+ {
289
+ provider: "openrouter",
290
+ model: "minimax/minimax-m2.7",
291
+ inputPerM: 0.3,
292
+ outputPerM: 1.2,
293
+ },
294
+ {
295
+ provider: "openrouter",
296
+ model: "minimax/minimax-m2",
297
+ inputPerM: 0.26,
298
+ outputPerM: 1,
299
+ },
70
300
  // OpenRouter — DeepSeek
71
- { provider: 'openrouter', model: 'deepseek/deepseek-v3.2-exp', inputPerM: 0.27, outputPerM: 0.41 },
72
- { provider: 'openrouter', model: 'deepseek/deepseek-r1', inputPerM: 0.70, outputPerM: 2.50 },
73
- { provider: 'openrouter', model: 'deepseek/deepseek-chat-v3.1', inputPerM: 0.15, outputPerM: 0.75 },
301
+ {
302
+ provider: "openrouter",
303
+ model: "deepseek/deepseek-v3.2-exp",
304
+ inputPerM: 0.27,
305
+ outputPerM: 0.41,
306
+ },
307
+ {
308
+ provider: "openrouter",
309
+ model: "deepseek/deepseek-r1",
310
+ inputPerM: 0.7,
311
+ outputPerM: 2.5,
312
+ },
313
+ {
314
+ provider: "openrouter",
315
+ model: "deepseek/deepseek-chat-v3.1",
316
+ inputPerM: 0.15,
317
+ outputPerM: 0.75,
318
+ },
74
319
  // OpenRouter — Qwen
75
- { provider: 'openrouter', model: 'qwen/qwen3-coder-plus', inputPerM: 0.65, outputPerM: 3.25 },
76
- { provider: 'openrouter', model: 'qwen/qwen3-235b-a22b-thinking-2507', inputPerM: 0.15, outputPerM: 1.50 },
77
- { provider: 'openrouter', model: 'qwen/qwen3-vl-235b-a22b-instruct', inputPerM: 0.20, outputPerM: 0.88 },
320
+ {
321
+ provider: "openrouter",
322
+ model: "qwen/qwen3-coder-plus",
323
+ inputPerM: 0.65,
324
+ outputPerM: 3.25,
325
+ },
326
+ {
327
+ provider: "openrouter",
328
+ model: "qwen/qwen3-235b-a22b-thinking-2507",
329
+ inputPerM: 0.15,
330
+ outputPerM: 1.5,
331
+ },
332
+ {
333
+ provider: "openrouter",
334
+ model: "qwen/qwen3-vl-235b-a22b-instruct",
335
+ inputPerM: 0.2,
336
+ outputPerM: 0.88,
337
+ },
78
338
  // OpenRouter — Tencent
79
- { provider: 'openrouter', model: 'tencent/hunyuan-a13b-instruct', inputPerM: 0.14, outputPerM: 0.57 },
339
+ {
340
+ provider: "openrouter",
341
+ model: "tencent/hunyuan-a13b-instruct",
342
+ inputPerM: 0.14,
343
+ outputPerM: 0.57,
344
+ },
80
345
  // OpenRouter — Z.ai (GLM)
81
- { provider: 'openrouter', model: 'z-ai/glm-5.1', inputPerM: 1.05, outputPerM: 3.50 },
82
- { provider: 'openrouter', model: 'z-ai/glm-5-turbo', inputPerM: 1.20, outputPerM: 4 },
83
- { provider: 'openrouter', model: 'z-ai/glm-5', inputPerM: 0.60, outputPerM: 2.08 },
84
- { provider: 'openrouter', model: 'z-ai/glm-4.7', inputPerM: 0.38, outputPerM: 1.74 },
85
- { provider: 'openrouter', model: 'z-ai/glm-4.7-flash', inputPerM: 0.06, outputPerM: 0.40 },
346
+ {
347
+ provider: "openrouter",
348
+ model: "z-ai/glm-5.1",
349
+ inputPerM: 1.05,
350
+ outputPerM: 3.5,
351
+ },
352
+ {
353
+ provider: "openrouter",
354
+ model: "z-ai/glm-5-turbo",
355
+ inputPerM: 1.2,
356
+ outputPerM: 4,
357
+ },
358
+ {
359
+ provider: "openrouter",
360
+ model: "z-ai/glm-5",
361
+ inputPerM: 0.6,
362
+ outputPerM: 2.08,
363
+ },
364
+ {
365
+ provider: "openrouter",
366
+ model: "z-ai/glm-4.7",
367
+ inputPerM: 0.38,
368
+ outputPerM: 1.74,
369
+ },
370
+ {
371
+ provider: "openrouter",
372
+ model: "z-ai/glm-4.7-flash",
373
+ inputPerM: 0.06,
374
+ outputPerM: 0.4,
375
+ },
86
376
  // OpenRouter — Free tier (2026-04-28 refresh — top-of-list per OR registry)
87
- { provider: 'openrouter', model: 'nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free', inputPerM: 0, outputPerM: 0, supportsThinking: true, family: 'nemotron' },
88
- { provider: 'openrouter', model: 'poolside/laguna-xs.2:free', inputPerM: 0, outputPerM: 0, family: 'poolside' },
89
- { provider: 'openrouter', model: 'poolside/laguna-m.1:free', inputPerM: 0, outputPerM: 0, family: 'poolside' },
90
- { provider: 'openrouter', model: 'inclusionai/ling-2.6-1t:free', inputPerM: 0, outputPerM: 0 },
91
- { provider: 'openrouter', model: 'tencent/hy3-preview:free', inputPerM: 0, outputPerM: 0 },
92
- { provider: 'openrouter', model: 'inclusionai/ling-2.6-flash:free', inputPerM: 0, outputPerM: 0 },
93
- { provider: 'openrouter', model: 'minimax/minimax-m2.5:free', inputPerM: 0, outputPerM: 0 },
94
- { provider: 'openrouter', model: 'qwen/qwen3-coder:free', inputPerM: 0, outputPerM: 0 },
95
- { provider: 'openrouter', model: 'z-ai/glm-4.5-air:free', inputPerM: 0, outputPerM: 0 },
96
- { provider: 'openrouter', model: 'google/gemma-4-31b-it:free', inputPerM: 0, outputPerM: 0 },
97
- { provider: 'openrouter', model: 'google/gemma-4-26b-a4b-it:free', inputPerM: 0, outputPerM: 0 },
377
+ {
378
+ provider: "openrouter",
379
+ model: "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free",
380
+ inputPerM: 0,
381
+ outputPerM: 0,
382
+ supportsThinking: true,
383
+ family: "nemotron",
384
+ },
385
+ {
386
+ provider: "openrouter",
387
+ model: "poolside/laguna-xs.2:free",
388
+ inputPerM: 0,
389
+ outputPerM: 0,
390
+ family: "poolside",
391
+ },
392
+ {
393
+ provider: "openrouter",
394
+ model: "poolside/laguna-m.1:free",
395
+ inputPerM: 0,
396
+ outputPerM: 0,
397
+ family: "poolside",
398
+ },
399
+ {
400
+ provider: "openrouter",
401
+ model: "inclusionai/ling-2.6-1t:free",
402
+ inputPerM: 0,
403
+ outputPerM: 0,
404
+ },
405
+ {
406
+ provider: "openrouter",
407
+ model: "tencent/hy3-preview:free",
408
+ inputPerM: 0,
409
+ outputPerM: 0,
410
+ },
411
+ {
412
+ provider: "openrouter",
413
+ model: "inclusionai/ling-2.6-flash:free",
414
+ inputPerM: 0,
415
+ outputPerM: 0,
416
+ },
417
+ {
418
+ provider: "openrouter",
419
+ model: "minimax/minimax-m2.5:free",
420
+ inputPerM: 0,
421
+ outputPerM: 0,
422
+ },
423
+ {
424
+ provider: "openrouter",
425
+ model: "qwen/qwen3-coder:free",
426
+ inputPerM: 0,
427
+ outputPerM: 0,
428
+ },
429
+ {
430
+ provider: "openrouter",
431
+ model: "z-ai/glm-4.5-air:free",
432
+ inputPerM: 0,
433
+ outputPerM: 0,
434
+ },
435
+ {
436
+ provider: "openrouter",
437
+ model: "google/gemma-4-31b-it:free",
438
+ inputPerM: 0,
439
+ outputPerM: 0,
440
+ },
441
+ {
442
+ provider: "openrouter",
443
+ model: "google/gemma-4-26b-a4b-it:free",
444
+ inputPerM: 0,
445
+ outputPerM: 0,
446
+ },
98
447
  // ─────────────────────────────────────────────────────────────────
99
448
  // Native providers added in 1.10.1 / catalogue refreshed 2026-04-28
100
449
  // ─────────────────────────────────────────────────────────────────
101
450
  // Mistral — native API. Latest tier as of 2026-04.
102
- { provider: 'mistral', model: 'mistral/mistral-large-latest', inputPerM: 2, outputPerM: 6, contextWindow: 128_000, family: 'mistral' },
103
- { provider: 'mistral', model: 'mistral/mistral-medium-latest', inputPerM: 0.4, outputPerM: 2, contextWindow: 128_000, family: 'mistral' },
104
- { provider: 'mistral', model: 'mistral/mistral-small-latest', inputPerM: 0.1, outputPerM: 0.3, contextWindow: 128_000, family: 'mistral' },
105
- { provider: 'mistral', model: 'mistral/codestral-latest', inputPerM: 0.3, outputPerM: 0.9, contextWindow: 256_000, family: 'mistral' },
106
- { provider: 'mistral', model: 'mistral/magistral-medium-latest', inputPerM: 2, outputPerM: 5, contextWindow: 128_000, supportsThinking: true, family: 'mistral' },
107
- { provider: 'mistral', model: 'mistral/ministral-8b-latest', inputPerM: 0.10, outputPerM: 0.10, contextWindow: 128_000, family: 'mistral' },
451
+ {
452
+ provider: "mistral",
453
+ model: "mistral/mistral-large-latest",
454
+ inputPerM: 2,
455
+ outputPerM: 6,
456
+ contextWindow: 128_000,
457
+ family: "mistral",
458
+ },
459
+ {
460
+ provider: "mistral",
461
+ model: "mistral/mistral-medium-latest",
462
+ inputPerM: 0.4,
463
+ outputPerM: 2,
464
+ contextWindow: 128_000,
465
+ family: "mistral",
466
+ },
467
+ {
468
+ provider: "mistral",
469
+ model: "mistral/mistral-small-latest",
470
+ inputPerM: 0.1,
471
+ outputPerM: 0.3,
472
+ contextWindow: 128_000,
473
+ family: "mistral",
474
+ },
475
+ {
476
+ provider: "mistral",
477
+ model: "mistral/codestral-latest",
478
+ inputPerM: 0.3,
479
+ outputPerM: 0.9,
480
+ contextWindow: 256_000,
481
+ family: "mistral",
482
+ },
483
+ {
484
+ provider: "mistral",
485
+ model: "mistral/magistral-medium-latest",
486
+ inputPerM: 2,
487
+ outputPerM: 5,
488
+ contextWindow: 128_000,
489
+ supportsThinking: true,
490
+ family: "mistral",
491
+ },
492
+ {
493
+ provider: "mistral",
494
+ model: "mistral/ministral-8b-latest",
495
+ inputPerM: 0.1,
496
+ outputPerM: 0.1,
497
+ contextWindow: 128_000,
498
+ family: "mistral",
499
+ },
108
500
  // Cohere — native API. command-a-* is the latest 2025 generation.
109
- { provider: 'cohere', model: 'cohere/command-a-03-2025', inputPerM: 2.5, outputPerM: 10, contextWindow: 256_000, family: 'cohere' },
110
- { provider: 'cohere', model: 'cohere/command-a-vision-07-2025', inputPerM: 2.5, outputPerM: 10, contextWindow: 256_000, family: 'cohere' },
111
- { provider: 'cohere', model: 'cohere/command-r-plus', inputPerM: 2.5, outputPerM: 10, contextWindow: 128_000, family: 'cohere' },
112
- { provider: 'cohere', model: 'cohere/command-r7b-12-2024', inputPerM: 0.0375, outputPerM: 0.15, contextWindow: 128_000, family: 'cohere' },
501
+ {
502
+ provider: "cohere",
503
+ model: "cohere/command-a-03-2025",
504
+ inputPerM: 2.5,
505
+ outputPerM: 10,
506
+ contextWindow: 256_000,
507
+ family: "cohere",
508
+ },
509
+ {
510
+ provider: "cohere",
511
+ model: "cohere/command-a-vision-07-2025",
512
+ inputPerM: 2.5,
513
+ outputPerM: 10,
514
+ contextWindow: 256_000,
515
+ family: "cohere",
516
+ },
517
+ {
518
+ provider: "cohere",
519
+ model: "cohere/command-r-plus",
520
+ inputPerM: 2.5,
521
+ outputPerM: 10,
522
+ contextWindow: 128_000,
523
+ family: "cohere",
524
+ },
525
+ {
526
+ provider: "cohere",
527
+ model: "cohere/command-r7b-12-2024",
528
+ inputPerM: 0.0375,
529
+ outputPerM: 0.15,
530
+ contextWindow: 128_000,
531
+ family: "cohere",
532
+ },
113
533
  // Cerebras — wafer-scale inference. Free during 2025 preview tier.
114
- { provider: 'cerebras', model: 'cerebras/qwen-3-235b-instruct', inputPerM: 0, outputPerM: 0, contextWindow: 65_000, family: 'qwen' },
115
- { provider: 'cerebras', model: 'cerebras/qwen-3-coder-480b', inputPerM: 0, outputPerM: 0, contextWindow: 256_000, family: 'qwen' },
116
- { provider: 'cerebras', model: 'cerebras/llama-4-maverick-17b-128e', inputPerM: 0, outputPerM: 0, contextWindow: 1_000_000, family: 'llama' },
117
- { provider: 'cerebras', model: 'cerebras/llama-3.3-70b', inputPerM: 0, outputPerM: 0, contextWindow: 128_000, family: 'llama' },
118
- { provider: 'cerebras', model: 'cerebras/gpt-oss-120b', inputPerM: 0, outputPerM: 0, contextWindow: 128_000, family: 'gpt-oss' },
534
+ {
535
+ provider: "cerebras",
536
+ model: "cerebras/qwen-3-235b-instruct",
537
+ inputPerM: 0,
538
+ outputPerM: 0,
539
+ contextWindow: 65_000,
540
+ family: "qwen",
541
+ },
542
+ {
543
+ provider: "cerebras",
544
+ model: "cerebras/qwen-3-coder-480b",
545
+ inputPerM: 0,
546
+ outputPerM: 0,
547
+ contextWindow: 256_000,
548
+ family: "qwen",
549
+ },
550
+ {
551
+ provider: "cerebras",
552
+ model: "cerebras/llama-4-maverick-17b-128e",
553
+ inputPerM: 0,
554
+ outputPerM: 0,
555
+ contextWindow: 1_000_000,
556
+ family: "llama",
557
+ },
558
+ {
559
+ provider: "cerebras",
560
+ model: "cerebras/llama-3.3-70b",
561
+ inputPerM: 0,
562
+ outputPerM: 0,
563
+ contextWindow: 128_000,
564
+ family: "llama",
565
+ },
566
+ {
567
+ provider: "cerebras",
568
+ model: "cerebras/gpt-oss-120b",
569
+ inputPerM: 0,
570
+ outputPerM: 0,
571
+ contextWindow: 128_000,
572
+ family: "gpt-oss",
573
+ },
119
574
  // Together AI — open-source model hub.
120
- { provider: 'together', model: 'meta-llama/Llama-3.3-70B-Instruct-Turbo', inputPerM: 0.88, outputPerM: 0.88, contextWindow: 128_000, family: 'llama' },
121
- { provider: 'together', model: 'deepseek-ai/DeepSeek-V3.1', inputPerM: 0.27, outputPerM: 1.10, contextWindow: 128_000, family: 'deepseek' },
122
- { provider: 'together', model: 'Qwen/Qwen3-235B-A22B-Instruct', inputPerM: 0.20, outputPerM: 0.60, contextWindow: 128_000, family: 'qwen' },
123
- { provider: 'together', model: 'Qwen/Qwen3-Coder-32B-Instruct', inputPerM: 0.18, outputPerM: 0.18, contextWindow: 128_000, family: 'qwen' },
124
- { provider: 'together', model: 'moonshotai/Kimi-K2-Instruct', inputPerM: 0.15, outputPerM: 0.60, contextWindow: 128_000, family: 'kimi' },
575
+ {
576
+ provider: "together",
577
+ model: "meta-llama/Llama-3.3-70B-Instruct-Turbo",
578
+ inputPerM: 0.88,
579
+ outputPerM: 0.88,
580
+ contextWindow: 128_000,
581
+ family: "llama",
582
+ },
583
+ {
584
+ provider: "together",
585
+ model: "deepseek-ai/DeepSeek-V3.1",
586
+ inputPerM: 0.27,
587
+ outputPerM: 1.1,
588
+ contextWindow: 128_000,
589
+ family: "deepseek",
590
+ },
591
+ {
592
+ provider: "together",
593
+ model: "Qwen/Qwen3-235B-A22B-Instruct",
594
+ inputPerM: 0.2,
595
+ outputPerM: 0.6,
596
+ contextWindow: 128_000,
597
+ family: "qwen",
598
+ },
599
+ {
600
+ provider: "together",
601
+ model: "Qwen/Qwen3-Coder-32B-Instruct",
602
+ inputPerM: 0.18,
603
+ outputPerM: 0.18,
604
+ contextWindow: 128_000,
605
+ family: "qwen",
606
+ },
607
+ {
608
+ provider: "together",
609
+ model: "moonshotai/Kimi-K2-Instruct",
610
+ inputPerM: 0.15,
611
+ outputPerM: 0.6,
612
+ contextWindow: 128_000,
613
+ family: "kimi",
614
+ },
125
615
  // Perplexity — Sonar (search-grounded). 5 USD per 1k searches included.
126
- { provider: 'perplexity', model: 'perplexity/sonar', inputPerM: 1, outputPerM: 1, contextWindow: 128_000, family: 'sonar' },
127
- { provider: 'perplexity', model: 'perplexity/sonar-pro', inputPerM: 3, outputPerM: 15, contextWindow: 200_000, family: 'sonar' },
128
- { provider: 'perplexity', model: 'perplexity/sonar-reasoning', inputPerM: 1, outputPerM: 5, contextWindow: 128_000, supportsThinking: true, family: 'sonar' },
129
- { provider: 'perplexity', model: 'perplexity/sonar-reasoning-pro', inputPerM: 2, outputPerM: 8, contextWindow: 200_000, supportsThinking: true, family: 'sonar' },
616
+ {
617
+ provider: "perplexity",
618
+ model: "perplexity/sonar",
619
+ inputPerM: 1,
620
+ outputPerM: 1,
621
+ contextWindow: 128_000,
622
+ family: "sonar",
623
+ },
624
+ {
625
+ provider: "perplexity",
626
+ model: "perplexity/sonar-pro",
627
+ inputPerM: 3,
628
+ outputPerM: 15,
629
+ contextWindow: 200_000,
630
+ family: "sonar",
631
+ },
632
+ {
633
+ provider: "perplexity",
634
+ model: "perplexity/sonar-reasoning",
635
+ inputPerM: 1,
636
+ outputPerM: 5,
637
+ contextWindow: 128_000,
638
+ supportsThinking: true,
639
+ family: "sonar",
640
+ },
641
+ {
642
+ provider: "perplexity",
643
+ model: "perplexity/sonar-reasoning-pro",
644
+ inputPerM: 2,
645
+ outputPerM: 8,
646
+ contextWindow: 200_000,
647
+ supportsThinking: true,
648
+ family: "sonar",
649
+ },
130
650
  // xAI (Grok) — native API. grok-4 is the current top reasoning model.
131
- { provider: 'xai', model: 'grok-4', inputPerM: 3, outputPerM: 15, contextWindow: 256_000, supportsThinking: true, family: 'grok' },
132
- { provider: 'xai', model: 'grok-4-fast', inputPerM: 0.20, outputPerM: 0.50, contextWindow: 2_000_000, family: 'grok' },
133
- { provider: 'xai', model: 'grok-code-fast-1', inputPerM: 0.20, outputPerM: 1.50, contextWindow: 256_000, family: 'grok' },
134
- { provider: 'xai', model: 'grok-3', inputPerM: 3, outputPerM: 15, contextWindow: 131_000, family: 'grok' },
651
+ {
652
+ provider: "xai",
653
+ model: "grok-4",
654
+ inputPerM: 3,
655
+ outputPerM: 15,
656
+ contextWindow: 256_000,
657
+ supportsThinking: true,
658
+ family: "grok",
659
+ },
660
+ {
661
+ provider: "xai",
662
+ model: "grok-4-fast",
663
+ inputPerM: 0.2,
664
+ outputPerM: 0.5,
665
+ contextWindow: 2_000_000,
666
+ family: "grok",
667
+ },
668
+ {
669
+ provider: "xai",
670
+ model: "grok-code-fast-1",
671
+ inputPerM: 0.2,
672
+ outputPerM: 1.5,
673
+ contextWindow: 256_000,
674
+ family: "grok",
675
+ },
676
+ {
677
+ provider: "xai",
678
+ model: "grok-3",
679
+ inputPerM: 3,
680
+ outputPerM: 15,
681
+ contextWindow: 131_000,
682
+ family: "grok",
683
+ },
135
684
  // Groq — LPU-accelerated inference. Free tier exists for most models.
136
- { provider: 'groq', model: 'groq/llama-3.3-70b-versatile', inputPerM: 0.59, outputPerM: 0.79, contextWindow: 128_000, family: 'llama' },
137
- { provider: 'groq', model: 'groq/llama-4-maverick-17b-128e-instruct', inputPerM: 0.20, outputPerM: 0.60, contextWindow: 1_000_000, family: 'llama' },
138
- { provider: 'groq', model: 'groq/llama-4-scout-17b-16e-instruct', inputPerM: 0.11, outputPerM: 0.34, contextWindow: 131_000, family: 'llama' },
139
- { provider: 'groq', model: 'groq/qwen-3-32b', inputPerM: 0.29, outputPerM: 0.59, contextWindow: 131_000, family: 'qwen' },
140
- { provider: 'groq', model: 'groq/moonshotai/kimi-k2-instruct-0905', inputPerM: 1, outputPerM: 3, contextWindow: 256_000, family: 'kimi' },
141
- { provider: 'groq', model: 'groq/openai/gpt-oss-120b', inputPerM: 0.15, outputPerM: 0.75, contextWindow: 131_000, family: 'gpt-oss' },
685
+ {
686
+ provider: "groq",
687
+ model: "groq/llama-3.3-70b-versatile",
688
+ inputPerM: 0.59,
689
+ outputPerM: 0.79,
690
+ contextWindow: 128_000,
691
+ family: "llama",
692
+ },
693
+ {
694
+ provider: "groq",
695
+ model: "groq/llama-4-maverick-17b-128e-instruct",
696
+ inputPerM: 0.2,
697
+ outputPerM: 0.6,
698
+ contextWindow: 1_000_000,
699
+ family: "llama",
700
+ },
701
+ {
702
+ provider: "groq",
703
+ model: "groq/llama-4-scout-17b-16e-instruct",
704
+ inputPerM: 0.11,
705
+ outputPerM: 0.34,
706
+ contextWindow: 131_000,
707
+ family: "llama",
708
+ },
709
+ {
710
+ provider: "groq",
711
+ model: "groq/qwen-3-32b",
712
+ inputPerM: 0.29,
713
+ outputPerM: 0.59,
714
+ contextWindow: 131_000,
715
+ family: "qwen",
716
+ },
717
+ {
718
+ provider: "groq",
719
+ model: "groq/moonshotai/kimi-k2-instruct-0905",
720
+ inputPerM: 1,
721
+ outputPerM: 3,
722
+ contextWindow: 256_000,
723
+ family: "kimi",
724
+ },
725
+ {
726
+ provider: "groq",
727
+ model: "groq/openai/gpt-oss-120b",
728
+ inputPerM: 0.15,
729
+ outputPerM: 0.75,
730
+ contextWindow: 131_000,
731
+ family: "gpt-oss",
732
+ },
142
733
  // Z.AI / GLM — native API. Lower latency than the OpenRouter mirror.
143
- { provider: 'zai', model: 'zai/glm-4.6', inputPerM: 0.60, outputPerM: 2.20, contextWindow: 200_000, family: 'glm' },
144
- { provider: 'zai', model: 'zai/glm-4.6-flash', inputPerM: 0.06, outputPerM: 0.40, contextWindow: 128_000, family: 'glm' },
145
- { provider: 'zai', model: 'zai/glm-4.5-air', inputPerM: 0, outputPerM: 0, contextWindow: 128_000, family: 'glm' },
734
+ {
735
+ provider: "zai",
736
+ model: "zai/glm-4.6",
737
+ inputPerM: 0.6,
738
+ outputPerM: 2.2,
739
+ contextWindow: 200_000,
740
+ family: "glm",
741
+ },
742
+ {
743
+ provider: "zai",
744
+ model: "zai/glm-4.6-flash",
745
+ inputPerM: 0.06,
746
+ outputPerM: 0.4,
747
+ contextWindow: 128_000,
748
+ family: "glm",
749
+ },
750
+ {
751
+ provider: "zai",
752
+ model: "zai/glm-4.5-air",
753
+ inputPerM: 0,
754
+ outputPerM: 0,
755
+ contextWindow: 128_000,
756
+ family: "glm",
757
+ },
146
758
  ];
147
759
  export function findPrice(provider, model) {
148
- return PRICES.find(p => p.provider === provider && p.model === model);
760
+ return PRICES.find((p) => p.provider === provider && p.model === model);
149
761
  }
150
762
  /**
151
763
  * Single-source-of-truth lookup. Returns the full PricePoint for a
@@ -154,7 +766,7 @@ export function findPrice(provider, model) {
154
766
  * a fallback default) when you only need the context limit.
155
767
  */
156
768
  export function lookupModel(modelId) {
157
- return PRICES.find(p => p.model === modelId);
769
+ return PRICES.find((p) => p.model === modelId);
158
770
  }
159
771
  /**
160
772
  * All models grouped by their `family` field (or by inferred family
@@ -174,28 +786,28 @@ export function modelsByFamily() {
174
786
  }
175
787
  function inferFamily(id) {
176
788
  if (/claude/.test(id))
177
- return 'claude';
789
+ return "claude";
178
790
  if (/^o[1-9]/.test(id) || /^gpt/.test(id) || /openai\//.test(id))
179
- return 'gpt';
791
+ return "gpt";
180
792
  if (/gemini/.test(id) || /google\//.test(id))
181
- return 'gemini';
793
+ return "gemini";
182
794
  if (/kimi/.test(id) || /moonshot/.test(id))
183
- return 'kimi';
795
+ return "kimi";
184
796
  if (/minimax/.test(id))
185
- return 'minimax';
797
+ return "minimax";
186
798
  if (/deepseek/.test(id))
187
- return 'deepseek';
799
+ return "deepseek";
188
800
  if (/qwen/.test(id))
189
- return 'qwen';
801
+ return "qwen";
190
802
  if (/llama|meta\//.test(id))
191
- return 'llama';
803
+ return "llama";
192
804
  if (/glm/.test(id) || /z-ai/.test(id))
193
- return 'glm';
805
+ return "glm";
194
806
  if (/hunyuan/.test(id) || /tencent\//.test(id))
195
- return 'hunyuan';
807
+ return "hunyuan";
196
808
  if (/ling/.test(id) || /inclusionai/.test(id))
197
- return 'ling';
198
- return 'other';
809
+ return "ling";
810
+ return "other";
199
811
  }
200
812
  /**
201
813
  * Best-effort lookup of the context-window cap for a model id. Searches
@@ -204,7 +816,7 @@ function inferFamily(id) {
204
816
  * (caller should fall back to a conservative default like 32k).
205
817
  */
206
818
  export function findContextWindow(modelId) {
207
- const hit = PRICES.find(p => p.model === modelId);
819
+ const hit = PRICES.find((p) => p.model === modelId);
208
820
  return hit?.contextWindow;
209
821
  }
210
822
  /**
@@ -217,62 +829,64 @@ export function findContextWindow(modelId) {
217
829
  export const DEFAULT_CONTEXT_WINDOW = 32_000;
218
830
  const CONTEXT_WINDOWS = {
219
831
  // Anthropic
220
- 'claude-opus-4-7': 200_000,
221
- 'claude-sonnet-4-6': 200_000,
222
- 'claude-haiku-4-5': 200_000,
832
+ "claude-opus-4-7": 200_000,
833
+ "claude-sonnet-4-6": 200_000,
834
+ "claude-haiku-4-5": 200_000,
223
835
  // OpenAI
224
- 'gpt-5.5-pro': 400_000,
225
- 'gpt-5.5': 400_000,
226
- 'gpt-5': 400_000,
227
- 'gpt-5-mini': 128_000,
228
- 'gpt-4o-mini': 128_000,
229
- 'o1': 200_000,
836
+ "gpt-5.5-pro": 400_000,
837
+ "gpt-5.5": 400_000,
838
+ "gpt-5": 400_000,
839
+ "gpt-5-mini": 128_000,
840
+ "gpt-4o-mini": 128_000,
841
+ o1: 200_000,
230
842
  // Gemini
231
- 'gemini-2.5-pro': 2_000_000,
232
- 'gemini-2.5-flash': 1_000_000,
843
+ "gemini-2.5-pro": 2_000_000,
844
+ "gemini-2.5-flash": 1_000_000,
233
845
  // NVIDIA NIM (verified live)
234
- 'deepseek-ai/deepseek-v4-pro': 128_000,
235
- 'deepseek-ai/deepseek-v4-flash': 128_000,
236
- 'moonshotai/kimi-k2-instruct': 128_000,
237
- 'qwen/qwen3-next-80b-a3b-instruct': 128_000,
238
- 'meta/llama-3.3-70b-instruct': 128_000,
846
+ "deepseek-ai/deepseek-v4-pro": 128_000,
847
+ "deepseek-ai/deepseek-v4-flash": 128_000,
848
+ "moonshotai/kimi-k2-instruct": 128_000,
849
+ "qwen/qwen3-next-80b-a3b-instruct": 128_000,
850
+ "meta/llama-3.3-70b-instruct": 128_000,
239
851
  // OpenRouter — frontier
240
- 'openai/gpt-5.5-pro': 400_000,
241
- 'openai/gpt-5.5': 400_000,
242
- 'google/gemini-3.1-pro-preview': 2_000_000,
243
- 'google/gemini-3-flash-preview': 1_000_000,
852
+ "openai/gpt-5.5-pro": 400_000,
853
+ "openai/gpt-5.5": 400_000,
854
+ "google/gemini-3.1-pro-preview": 2_000_000,
855
+ "google/gemini-3-flash-preview": 1_000_000,
244
856
  // OpenRouter — Moonshot/Kimi
245
- 'moonshotai/kimi-k2.6': 128_000,
246
- 'moonshotai/kimi-k2.5': 128_000,
247
- 'moonshotai/kimi-k2-thinking': 128_000,
857
+ "moonshotai/kimi-k2.6": 128_000,
858
+ "moonshotai/kimi-k2.5": 128_000,
859
+ "moonshotai/kimi-k2-thinking": 128_000,
248
860
  // OpenRouter — MiniMax
249
- 'minimax/minimax-m2.7': 200_000,
250
- 'minimax/minimax-m2': 200_000,
861
+ "minimax/minimax-m2.7": 200_000,
862
+ "minimax/minimax-m2": 200_000,
251
863
  // OpenRouter — DeepSeek
252
- 'deepseek/deepseek-v3.2-exp': 128_000,
253
- 'deepseek/deepseek-r1': 128_000,
254
- 'deepseek/deepseek-chat-v3.1': 128_000,
864
+ "deepseek/deepseek-v3.2-exp": 128_000,
865
+ "deepseek/deepseek-r1": 128_000,
866
+ "deepseek/deepseek-chat-v3.1": 128_000,
255
867
  // OpenRouter — Qwen
256
- 'qwen/qwen3-coder-plus': 256_000,
257
- 'qwen/qwen3-235b-a22b-thinking-2507': 128_000,
258
- 'qwen/qwen3-vl-235b-a22b-instruct': 128_000,
868
+ "qwen/qwen3-coder-plus": 256_000,
869
+ "qwen/qwen3-235b-a22b-thinking-2507": 128_000,
870
+ "qwen/qwen3-vl-235b-a22b-instruct": 128_000,
259
871
  // OpenRouter — Tencent
260
- 'tencent/hunyuan-a13b-instruct': 32_000,
872
+ "tencent/hunyuan-a13b-instruct": 32_000,
261
873
  // OpenRouter — Z.ai (GLM)
262
- 'z-ai/glm-5.1': 128_000,
263
- 'z-ai/glm-5-turbo': 128_000,
264
- 'z-ai/glm-5': 128_000,
265
- 'z-ai/glm-4.7': 128_000,
266
- 'z-ai/glm-4.7-flash': 32_000,
874
+ "z-ai/glm-5.1": 128_000,
875
+ "z-ai/glm-5-turbo": 128_000,
876
+ "z-ai/glm-5": 128_000,
877
+ "z-ai/glm-4.7": 128_000,
878
+ "z-ai/glm-4.7-flash": 32_000,
267
879
  // OpenRouter — free
268
- 'inclusionai/ling-2.6-1t:free': 32_000,
269
- 'tencent/hy3-preview:free': 32_000,
270
- 'minimax/minimax-m2.5:free': 200_000,
271
- 'qwen/qwen3-coder:free': 32_000,
272
- 'z-ai/glm-4.5-air:free': 128_000,
880
+ "inclusionai/ling-2.6-1t:free": 32_000,
881
+ "tencent/hy3-preview:free": 32_000,
882
+ "minimax/minimax-m2.5:free": 200_000,
883
+ "qwen/qwen3-coder:free": 32_000,
884
+ "z-ai/glm-4.5-air:free": 128_000,
273
885
  };
274
886
  export function contextWindowFor(modelId) {
275
- return CONTEXT_WINDOWS[modelId] ?? findContextWindow(modelId) ?? DEFAULT_CONTEXT_WINDOW;
887
+ return (CONTEXT_WINDOWS[modelId] ??
888
+ findContextWindow(modelId) ??
889
+ DEFAULT_CONTEXT_WINDOW);
276
890
  }
277
891
  /**
278
892
  * Failover map: when the primary model is unavailable (provider 504,
@@ -289,18 +903,27 @@ export function contextWindowFor(modelId) {
289
903
  */
290
904
  const MODEL_FAILOVERS = {
291
905
  // NIM → OpenRouter mirror
292
- 'moonshotai/kimi-k2-instruct': 'moonshotai/kimi-k2.5',
293
- 'qwen/qwen3-next-80b-a3b-instruct': 'qwen/qwen3-235b-a22b-thinking-2507',
294
- 'meta/llama-3.3-70b-instruct': 'qwen/qwen3-coder:free',
295
- 'deepseek-ai/deepseek-v4-pro': 'deepseek/deepseek-v3.2-exp',
296
- 'deepseek-ai/deepseek-v4-flash': 'deepseek/deepseek-chat-v3.1',
906
+ "moonshotai/kimi-k2-instruct": "moonshotai/kimi-k2.5",
907
+ "qwen/qwen3-next-80b-a3b-instruct": "qwen/qwen3-235b-a22b-thinking-2507",
908
+ "meta/llama-3.3-70b-instruct": "qwen/qwen3-coder:free",
909
+ "deepseek-ai/deepseek-v4-pro": "deepseek/deepseek-v3.2-exp",
910
+ "deepseek-ai/deepseek-v4-flash": "deepseek/deepseek-chat-v3.1",
297
911
  // Anthropic-native → OpenRouter mirror (no key needed if OR set)
298
- 'claude-opus-4-7': 'anthropic/claude-opus-4-7',
299
- 'claude-sonnet-4-6': 'anthropic/claude-sonnet-4-6',
300
- 'claude-haiku-4-5': 'anthropic/claude-haiku-4-5',
912
+ "claude-opus-4-7": "anthropic/claude-opus-4-7",
913
+ "claude-sonnet-4-6": "anthropic/claude-sonnet-4-6",
914
+ "claude-haiku-4-5": "anthropic/claude-haiku-4-5",
301
915
  };
916
+ import { familyAlternatives } from "../providers/family-fallback.js";
302
917
  export function findFailover(modelId) {
303
- return MODEL_FAILOVERS[modelId];
918
+ // First try exact failover map
919
+ const exact = MODEL_FAILOVERS[modelId];
920
+ if (exact)
921
+ return exact;
922
+ // Then try family-aware alternatives
923
+ const fam = familyAlternatives(modelId);
924
+ if (fam.length > 0)
925
+ return fam[0].model;
926
+ return undefined;
304
927
  }
305
928
  /**
306
929
  * Model IDs the upstream provider has dropped (returns 400 "not a
@@ -329,28 +952,28 @@ export function migrateDeprecatedModel(modelId) {
329
952
  */
330
953
  const MODEL_ALIASES = {
331
954
  // Anthropic
332
- opus: 'claude-opus-4-7',
333
- sonnet: 'claude-sonnet-4-6',
334
- haiku: 'claude-haiku-4-5',
955
+ opus: "claude-opus-4-7",
956
+ sonnet: "claude-sonnet-4-6",
957
+ haiku: "claude-haiku-4-5",
335
958
  // OpenAI
336
- gpt5: 'gpt-5',
337
- 'gpt-5': 'gpt-5',
338
- 'gpt5-pro': 'gpt-5.5-pro',
339
- 'gpt5-mini': 'gpt-5-mini',
340
- o1: 'o1',
959
+ gpt5: "gpt-5",
960
+ "gpt-5": "gpt-5",
961
+ "gpt5-pro": "gpt-5.5-pro",
962
+ "gpt5-mini": "gpt-5-mini",
963
+ o1: "o1",
341
964
  // Gemini
342
- gemini: 'gemini-2.5-pro',
343
- flash: 'gemini-2.5-flash',
965
+ gemini: "gemini-2.5-pro",
966
+ flash: "gemini-2.5-flash",
344
967
  // NVIDIA NIM
345
- kimi: 'moonshotai/kimi-k2-instruct',
346
- qwen: 'qwen/qwen3-next-80b-a3b-instruct',
347
- llama: 'meta/llama-3.3-70b-instruct',
348
- deepseek: 'deepseek-ai/deepseek-v4-pro',
349
- 'deepseek-pro': 'deepseek-ai/deepseek-v4-pro',
350
- 'deepseek-flash': 'deepseek-ai/deepseek-v4-flash',
968
+ kimi: "moonshotai/kimi-k2-instruct",
969
+ qwen: "qwen/qwen3-next-80b-a3b-instruct",
970
+ llama: "meta/llama-3.3-70b-instruct",
971
+ deepseek: "deepseek-ai/deepseek-v4-pro",
972
+ "deepseek-pro": "deepseek-ai/deepseek-v4-pro",
973
+ "deepseek-flash": "deepseek-ai/deepseek-v4-flash",
351
974
  // OpenRouter free tier
352
- ling: 'inclusionai/ling-2.6-1t:free',
353
- hy3: 'tencent/hy3-preview:free',
975
+ ling: "inclusionai/ling-2.6-1t:free",
976
+ hy3: "tencent/hy3-preview:free",
354
977
  };
355
978
  export function resolveModelAlias(input) {
356
979
  if (!input)
@@ -359,6 +982,9 @@ export function resolveModelAlias(input) {
359
982
  return MODEL_ALIASES[key] ?? input;
360
983
  }
361
984
  export function listModelAliases() {
362
- return Object.entries(MODEL_ALIASES).map(([alias, model]) => ({ alias, model }));
985
+ return Object.entries(MODEL_ALIASES).map(([alias, model]) => ({
986
+ alias,
987
+ model,
988
+ }));
363
989
  }
364
990
  //# sourceMappingURL=prices.js.map