@omnicross/contracts 0.3.0 → 0.4.0

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 (31) hide show
  1. package/dist/account-tokens-types.d.cts +76 -1
  2. package/dist/account-tokens-types.d.ts +76 -1
  3. package/dist/canonical-models.cjs +49 -4
  4. package/dist/canonical-models.d.cts +1 -1
  5. package/dist/canonical-models.d.ts +1 -1
  6. package/dist/canonical-models.js +49 -4
  7. package/dist/endpoint-resolver.d.cts +1 -1
  8. package/dist/endpoint-resolver.d.ts +1 -1
  9. package/dist/index.cjs +328 -6
  10. package/dist/index.d.cts +2 -2
  11. package/dist/index.d.ts +2 -2
  12. package/dist/index.js +328 -6
  13. package/dist/{llm-config-DeWNx1ig.d.cts → llm-config-CEZvId7g.d.cts} +18 -0
  14. package/dist/{llm-config-CKOaFFdy.d.ts → llm-config-Dxw7F1_U.d.ts} +18 -0
  15. package/dist/llm-config.d.cts +1 -1
  16. package/dist/llm-config.d.ts +1 -1
  17. package/dist/provider-presets/index.cjs +209 -0
  18. package/dist/provider-presets/index.d.cts +2 -2
  19. package/dist/provider-presets/index.d.ts +2 -2
  20. package/dist/provider-presets/index.js +209 -0
  21. package/dist/subscription-model-catalog.cjs +70 -2
  22. package/dist/subscription-model-catalog.d.cts +1 -1
  23. package/dist/subscription-model-catalog.d.ts +1 -1
  24. package/dist/subscription-model-catalog.js +70 -2
  25. package/dist/subscription-types.d.cts +1 -1
  26. package/dist/subscription-types.d.ts +1 -1
  27. package/dist/thinking-config.cjs +49 -4
  28. package/dist/thinking-config.d.cts +1 -1
  29. package/dist/thinking-config.d.ts +1 -1
  30. package/dist/thinking-config.js +49 -4
  31. package/package.json +1 -1
package/dist/account-tokens-types.d.cts CHANGED
@@ -122,6 +122,72 @@ type GeminiTokenConfig = {
122
122
  lastRefreshedAt?: string;
123
123
  errorMessage?: string;
124
124
  };
125
+ /**
126
+ * Kimi Code (Moonshot) token configuration. OAuth is a RFC 8628 device flow
127
+ * (`auth.kimi.com`); the access token serves BOTH inference
128
+ * (`api.kimi.com/coding/v1`, anthropic + openai faces) and the
129
+ * `/coding/v1/usages` quota endpoint. `accountId` is the access-token JWT's
130
+ * `user_id | sub` claim; `deviceId` backs the `X-Msh-Device-Id` fingerprint
131
+ * header Kimi expects (stable per account, minted at login).
132
+ */
133
+ type KimiTokenConfig = {
134
+ authMethod: AuthMethod;
135
+ status: TokenStatus;
136
+ accessToken?: string;
137
+ refreshToken?: string;
138
+ expiresAt?: string;
139
+ accountId?: string;
140
+ deviceId?: string;
141
+ lastRefreshedAt?: string;
142
+ errorMessage?: string;
143
+ /** Managed-account credential warning (duplicate-token, projected on listing). */
144
+ syncWarning?: SyncWarningCode;
145
+ };
146
+ /**
147
+ * Grok (xAI SuperGrok) token configuration. OAuth is a RFC 8628 device flow
148
+ * at `auth.x.ai` (token endpoint resolved + host-pinned via OIDC discovery);
149
+ * the access token serves BOTH inference (`api.x.ai/v1/responses`) and the
150
+ * `cli-chat-proxy.grok.com/v1/billing` quota endpoint. `accountId` is the
151
+ * access-token JWT's `sub` claim.
152
+ */
153
+ type GrokTokenConfig = {
154
+ authMethod: AuthMethod;
155
+ status: TokenStatus;
156
+ accessToken?: string;
157
+ refreshToken?: string;
158
+ expiresAt?: string;
159
+ accountId?: string;
160
+ lastRefreshedAt?: string;
161
+ errorMessage?: string;
162
+ /** Managed-account credential warning (duplicate-token, projected on listing). */
163
+ syncWarning?: SyncWarningCode;
164
+ };
165
+ /**
166
+ * GitHub Copilot token configuration. OAuth is a RFC 8628 device flow at
167
+ * `github.com` (official Copilot CLI app, scope `read:user`); the minted
168
+ * `ghu_` token is LONG-LIVED — refresh is a LOCAL no-op (access and refresh
169
+ * are the same token, far-future expiry), so no refresh network call exists.
170
+ * Inference rides the Copilot API (`api.githubcopilot.com` by default; the
171
+ * plan-advertised `apiEndpoint` discovered at login wins); the quota endpoint
172
+ * is `api.github.com/copilot_internal/user`. `accountId` is the GitHub login.
173
+ */
174
+ type CopilotTokenConfig = {
175
+ authMethod: AuthMethod;
176
+ status: TokenStatus;
177
+ accessToken?: string;
178
+ refreshToken?: string;
179
+ expiresAt?: string;
180
+ accountId?: string;
181
+ email?: string;
182
+ /** Plan-advertised Copilot API endpoint (from `copilot_internal/user.endpoints.api`). */
183
+ apiEndpoint?: string;
184
+ /** GitHub Enterprise domain (e.g. `company.ghe.com`); personal accounts leave it unset. */
185
+ enterpriseUrl?: string;
186
+ lastRefreshedAt?: string;
187
+ errorMessage?: string;
188
+ /** Managed-account credential warning (duplicate-token, projected on listing). */
189
+ syncWarning?: SyncWarningCode;
190
+ };
125
191
  /**
126
192
  * A single subscription account entry. The provider's existing token config
127
193
  * is carried verbatim under a nested `tokens` field; entry metadata
@@ -210,6 +276,9 @@ type AccountTokensConfig = {
210
276
  codex?: CodexTokenConfig;
211
277
  gemini?: GeminiTokenConfig;
212
278
  opencodego?: OpenCodeGoTokenConfig;
279
+ kimi?: KimiTokenConfig;
280
+ grok?: GrokTokenConfig;
281
+ copilot?: CopilotTokenConfig;
213
282
  claudeAccounts?: SubscriptionAccountEntry<ClaudeTokenConfig>[];
214
283
  activeClaudeAccountId?: string;
215
284
  codexAccounts?: SubscriptionAccountEntry<CodexTokenConfig>[];
@@ -218,6 +287,12 @@ type AccountTokensConfig = {
218
287
  activeGeminiAccountId?: string;
219
288
  opencodegoAccounts?: SubscriptionAccountEntry<OpenCodeGoTokenConfig>[];
220
289
  activeOpencodegoAccountId?: string;
290
+ kimiAccounts?: SubscriptionAccountEntry<KimiTokenConfig>[];
291
+ activeKimiAccountId?: string;
292
+ grokAccounts?: SubscriptionAccountEntry<GrokTokenConfig>[];
293
+ activeGrokAccountId?: string;
294
+ copilotAccounts?: SubscriptionAccountEntry<CopilotTokenConfig>[];
295
+ activeCopilotAccountId?: string;
221
296
  updatedAt: string;
222
297
  };
223
298
  /**
@@ -340,4 +415,4 @@ type TokenExchangeRequest = {
340
415
  state: string;
341
416
  };
342
417
 
343
- export type { AccountClientIdentity, AccountTokensConfig, AuthMethod, ClaudeAuthMethod, ClaudeTokenConfig, CodexTokenConfig, GeminiTokenConfig, OAuthParams, ProxyConfig, SanitizedProxyConfig, SubscriptionAccountEntry, SubscriptionAccountSanitized, SubscriptionLevel, SyncWarningCode, TokenExchangeRequest, TokenStatus };
418
+ export type { AccountClientIdentity, AccountTokensConfig, AuthMethod, ClaudeAuthMethod, ClaudeTokenConfig, CodexTokenConfig, CopilotTokenConfig, GeminiTokenConfig, GrokTokenConfig, KimiTokenConfig, OAuthParams, ProxyConfig, SanitizedProxyConfig, SubscriptionAccountEntry, SubscriptionAccountSanitized, SubscriptionLevel, SyncWarningCode, TokenExchangeRequest, TokenStatus };
package/dist/account-tokens-types.d.ts CHANGED
@@ -122,6 +122,72 @@ type GeminiTokenConfig = {
122
122
  lastRefreshedAt?: string;
123
123
  errorMessage?: string;
124
124
  };
125
+ /**
126
+ * Kimi Code (Moonshot) token configuration. OAuth is a RFC 8628 device flow
127
+ * (`auth.kimi.com`); the access token serves BOTH inference
128
+ * (`api.kimi.com/coding/v1`, anthropic + openai faces) and the
129
+ * `/coding/v1/usages` quota endpoint. `accountId` is the access-token JWT's
130
+ * `user_id | sub` claim; `deviceId` backs the `X-Msh-Device-Id` fingerprint
131
+ * header Kimi expects (stable per account, minted at login).
132
+ */
133
+ type KimiTokenConfig = {
134
+ authMethod: AuthMethod;
135
+ status: TokenStatus;
136
+ accessToken?: string;
137
+ refreshToken?: string;
138
+ expiresAt?: string;
139
+ accountId?: string;
140
+ deviceId?: string;
141
+ lastRefreshedAt?: string;
142
+ errorMessage?: string;
143
+ /** Managed-account credential warning (duplicate-token, projected on listing). */
144
+ syncWarning?: SyncWarningCode;
145
+ };
146
+ /**
147
+ * Grok (xAI SuperGrok) token configuration. OAuth is a RFC 8628 device flow
148
+ * at `auth.x.ai` (token endpoint resolved + host-pinned via OIDC discovery);
149
+ * the access token serves BOTH inference (`api.x.ai/v1/responses`) and the
150
+ * `cli-chat-proxy.grok.com/v1/billing` quota endpoint. `accountId` is the
151
+ * access-token JWT's `sub` claim.
152
+ */
153
+ type GrokTokenConfig = {
154
+ authMethod: AuthMethod;
155
+ status: TokenStatus;
156
+ accessToken?: string;
157
+ refreshToken?: string;
158
+ expiresAt?: string;
159
+ accountId?: string;
160
+ lastRefreshedAt?: string;
161
+ errorMessage?: string;
162
+ /** Managed-account credential warning (duplicate-token, projected on listing). */
163
+ syncWarning?: SyncWarningCode;
164
+ };
165
+ /**
166
+ * GitHub Copilot token configuration. OAuth is a RFC 8628 device flow at
167
+ * `github.com` (official Copilot CLI app, scope `read:user`); the minted
168
+ * `ghu_` token is LONG-LIVED — refresh is a LOCAL no-op (access and refresh
169
+ * are the same token, far-future expiry), so no refresh network call exists.
170
+ * Inference rides the Copilot API (`api.githubcopilot.com` by default; the
171
+ * plan-advertised `apiEndpoint` discovered at login wins); the quota endpoint
172
+ * is `api.github.com/copilot_internal/user`. `accountId` is the GitHub login.
173
+ */
174
+ type CopilotTokenConfig = {
175
+ authMethod: AuthMethod;
176
+ status: TokenStatus;
177
+ accessToken?: string;
178
+ refreshToken?: string;
179
+ expiresAt?: string;
180
+ accountId?: string;
181
+ email?: string;
182
+ /** Plan-advertised Copilot API endpoint (from `copilot_internal/user.endpoints.api`). */
183
+ apiEndpoint?: string;
184
+ /** GitHub Enterprise domain (e.g. `company.ghe.com`); personal accounts leave it unset. */
185
+ enterpriseUrl?: string;
186
+ lastRefreshedAt?: string;
187
+ errorMessage?: string;
188
+ /** Managed-account credential warning (duplicate-token, projected on listing). */
189
+ syncWarning?: SyncWarningCode;
190
+ };
125
191
  /**
126
192
  * A single subscription account entry. The provider's existing token config
127
193
  * is carried verbatim under a nested `tokens` field; entry metadata
@@ -210,6 +276,9 @@ type AccountTokensConfig = {
210
276
  codex?: CodexTokenConfig;
211
277
  gemini?: GeminiTokenConfig;
212
278
  opencodego?: OpenCodeGoTokenConfig;
279
+ kimi?: KimiTokenConfig;
280
+ grok?: GrokTokenConfig;
281
+ copilot?: CopilotTokenConfig;
213
282
  claudeAccounts?: SubscriptionAccountEntry<ClaudeTokenConfig>[];
214
283
  activeClaudeAccountId?: string;
215
284
  codexAccounts?: SubscriptionAccountEntry<CodexTokenConfig>[];
@@ -218,6 +287,12 @@ type AccountTokensConfig = {
218
287
  activeGeminiAccountId?: string;
219
288
  opencodegoAccounts?: SubscriptionAccountEntry<OpenCodeGoTokenConfig>[];
220
289
  activeOpencodegoAccountId?: string;
290
+ kimiAccounts?: SubscriptionAccountEntry<KimiTokenConfig>[];
291
+ activeKimiAccountId?: string;
292
+ grokAccounts?: SubscriptionAccountEntry<GrokTokenConfig>[];
293
+ activeGrokAccountId?: string;
294
+ copilotAccounts?: SubscriptionAccountEntry<CopilotTokenConfig>[];
295
+ activeCopilotAccountId?: string;
221
296
  updatedAt: string;
222
297
  };
223
298
  /**
@@ -340,4 +415,4 @@ type TokenExchangeRequest = {
340
415
  state: string;
341
416
  };
342
417
 
343
- export type { AccountClientIdentity, AccountTokensConfig, AuthMethod, ClaudeAuthMethod, ClaudeTokenConfig, CodexTokenConfig, GeminiTokenConfig, OAuthParams, ProxyConfig, SanitizedProxyConfig, SubscriptionAccountEntry, SubscriptionAccountSanitized, SubscriptionLevel, SyncWarningCode, TokenExchangeRequest, TokenStatus };
418
+ export type { AccountClientIdentity, AccountTokensConfig, AuthMethod, ClaudeAuthMethod, ClaudeTokenConfig, CodexTokenConfig, CopilotTokenConfig, GeminiTokenConfig, GrokTokenConfig, KimiTokenConfig, OAuthParams, ProxyConfig, SanitizedProxyConfig, SubscriptionAccountEntry, SubscriptionAccountSanitized, SubscriptionLevel, SyncWarningCode, TokenExchangeRequest, TokenStatus };
package/dist/canonical-models.cjs CHANGED
@@ -55,19 +55,56 @@ var ANTHROPIC_MODELS = {
55
55
  "claude-opus-4-6[1m]": { category: "reasoning", contextLength: 1e6, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 128e3 } },
56
56
  "claude-sonnet-4-6": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } },
57
57
  "claude-sonnet-4-6[1m]": { category: "chat", contextLength: 1e6, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } },
58
- "claude-haiku-4-5": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } }
58
+ "claude-haiku-4-5": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } },
59
+ // Copilot-served variants (same model family on the Copilot API; context
60
+ // windows per the Copilot catalog where it diverges from the family default).
61
+ "claude-opus-4-5": { category: "reasoning", contextLength: 2e5, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true },
62
+ "claude-sonnet-4-5": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true },
63
+ "claude-sonnet-4": { category: "chat", contextLength: 216e3, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true }
59
64
  };
60
65
  var GEMINI_MODELS = {
61
66
  "gemini-3.5-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 24576 } },
62
67
  "gemini-3-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 24576 } },
63
- "gemini-3.1-pro": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 128, max: 32768 } }
68
+ "gemini-3.1-pro": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 128, max: 32768 } },
69
+ // Copilot-served variants (chat-completions wire on the Copilot API).
70
+ "gemini-3.1-pro-preview": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"] },
71
+ "gemini-3.6-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"] },
72
+ "gemini-3.7-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"] },
73
+ "gemini-3-pro-preview": { category: "chat", contextLength: 128e3, maxTokens: 65536, vision: true, functionCall: true, reasoning: true },
74
+ "gemini-3-flash-preview": { category: "chat", contextLength: 128e3, maxTokens: 65536, vision: true, functionCall: true, reasoning: true },
75
+ "gemini-2.5-pro": { category: "chat", contextLength: 128e3, maxTokens: 65536, vision: true, functionCall: true, reasoning: true }
64
76
  };
65
77
  var GROK_MODELS = {
66
78
  "grok-4.5": { category: "chat", contextLength: 5e5, maxTokens: 32768, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
67
79
  "grok-4.3": { category: "chat", contextLength: 1e6, maxTokens: 32768, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
68
80
  "grok-4.20": { category: "chat", contextLength: 131072, maxTokens: 32768, vision: true, functionCall: true },
69
81
  "grok-4.20-reasoning": { category: "reasoning", contextLength: 131072, maxTokens: 32768, reasoning: true, vision: true, functionCall: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
70
- "grok-4.20-multi-agent": { category: "reasoning", contextLength: 131072, maxTokens: 32768, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } }
82
+ "grok-4.20-multi-agent": { category: "reasoning", contextLength: 131072, maxTokens: 32768, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
83
+ // SuperGrok subscription ids (Responses wire). Upstream context/output caps
84
+ // per the audit source; the effort-taking models accept low/medium/high only
85
+ // (minimal→low, xhigh/max→high), and the omit-effort models carry
86
+ // `thinkingLevels: ['none']` so the effort negotiation drops the
87
+ // `reasoning.effort` field entirely (sending it upstream is a 400).
88
+ "grok-4.6": { category: "chat", contextLength: 5e5, maxTokens: 5e5, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
89
+ "grok-4.20-0309-non-reasoning": { category: "chat", contextLength: 2e6, maxTokens: 2e6, vision: true, functionCall: true, thinkingLevels: ["none"] },
90
+ "grok-4.20-0309-reasoning": { category: "reasoning", contextLength: 2e6, maxTokens: 2e6, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none"] },
91
+ "grok-4.20-multi-agent-0309": { category: "reasoning", contextLength: 2e6, maxTokens: 2e6, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
92
+ "grok-build": { category: "reasoning", contextLength: 512e3, maxTokens: 512e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none"] },
93
+ "grok-build-0.1": { category: "reasoning", contextLength: 256e3, maxTokens: 256e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none"] },
94
+ "grok-composer-2.5-fast": { category: "chat", contextLength: 2e5, maxTokens: 2e5, functionCall: true, thinkingLevels: ["none"] }
95
+ };
96
+ var COPILOT_MODELS = {
97
+ "gpt-5.1-codex-max": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
98
+ "gpt-5.1-codex-mini": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
99
+ "gpt-5.2-codex": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
100
+ "gpt-5.1-codex": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
101
+ "gpt-5.1": { category: "reasoning", contextLength: 264e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
102
+ "gpt-5-mini": { category: "reasoning", contextLength: 264e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
103
+ "gpt-5.4-nano": { category: "reasoning", contextLength: 4e5, maxTokens: 128e3, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
104
+ "grok-code-fast-1": { category: "code", contextLength: 192e3, maxTokens: 32768, functionCall: true, reasoning: true },
105
+ "raptor-mini": { category: "reasoning", contextLength: 4e5, maxTokens: 65536, functionCall: true, reasoning: true, vision: true },
106
+ "mai-code-1-flash-picker": { category: "code", contextLength: 256e3, maxTokens: 65536, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"] },
107
+ "mai-code-1.1-flash": { category: "code", contextLength: 256e3, maxTokens: 65536, functionCall: true, reasoning: true, vision: true, thinkingLevels: ["none", "low", "medium", "high"] }
71
108
  };
72
109
  var DEEPSEEK_MODELS = {
73
110
  "deepseek-v4-flash": { category: "chat", contextLength: 1e6, maxTokens: 384e3, functionCall: true, reasoning: true, thinkingLevels: ["none", "high", "max"], thinkingTokenLimit: { min: 0, max: 32768 } },
@@ -77,6 +114,8 @@ var DEEPSEEK_MODELS = {
77
114
  "deepseek-r1": { category: "reasoning", contextLength: 64e3, maxTokens: 8192, reasoning: true, thinkingTokenLimit: { min: 0, max: 32768 } }
78
115
  };
79
116
  var ZHIPU_MODELS = {
117
+ "glm-5.3": { category: "chat", contextLength: 1048576, maxTokens: 131072, functionCall: true },
118
+ "glm-5.3-flash": { category: "chat", contextLength: 1048576, maxTokens: 131072, functionCall: true },
80
119
  "glm-5.2": { category: "chat", contextLength: 1048576, maxTokens: 131072, functionCall: true },
81
120
  "glm-5": { category: "chat", contextLength: 2e5, maxTokens: 128e3, functionCall: true },
82
121
  "glm-5.1": { category: "chat", contextLength: 2e5, maxTokens: 128e3, functionCall: true },
@@ -111,10 +150,15 @@ var TENCENT_HUNYUAN_MODELS = {
111
150
  "hunyuan-turbos": { category: "chat", functionCall: true }
112
151
  };
113
152
  var MINIMAX_MODELS = {
153
+ "minimax-m3": { category: "code", contextLength: 204800, functionCall: true, reasoning: true },
154
+ "minimax-m2.7": { category: "code", contextLength: 204800, functionCall: true },
155
+ "minimax-m2.7-highspeed": { category: "code", contextLength: 204800, functionCall: true },
114
156
  "minimax-m2.5": { category: "code", contextLength: 204800, functionCall: true },
115
157
  "minimax-m2.5-highspeed": { category: "code", contextLength: 204800, functionCall: true },
158
+ "minimax-m2.5-lightning": { category: "code", contextLength: 204800, functionCall: true },
116
159
  "minimax-m2.1": { category: "code", contextLength: 204800, functionCall: true },
117
- "minimax-m2.1-highspeed": { category: "code", contextLength: 204800, functionCall: true }
160
+ "minimax-m2.1-lightning": { category: "code", contextLength: 204800, functionCall: true },
161
+ "minimax-m2": { category: "code", contextLength: 204800, functionCall: true }
118
162
  };
119
163
  var XIAOMI_MIMO_CANONICAL = {
120
164
  "mimo-v2.5": { category: "code", contextLength: 1e6, maxTokens: 32768, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
@@ -156,6 +200,7 @@ var VENDOR_GROUPS = [
156
200
  { name: "ANTHROPIC_MODELS", entries: ANTHROPIC_MODELS },
157
201
  { name: "GEMINI_MODELS", entries: GEMINI_MODELS },
158
202
  { name: "GROK_MODELS", entries: GROK_MODELS },
203
+ { name: "COPILOT_MODELS", entries: COPILOT_MODELS },
159
204
  { name: "DEEPSEEK_MODELS", entries: DEEPSEEK_MODELS },
160
205
  { name: "ZHIPU_MODELS", entries: ZHIPU_MODELS },
161
206
  { name: "KIMI_MODELS", entries: KIMI_MODELS },
package/dist/canonical-models.d.cts CHANGED
@@ -1,5 +1,5 @@
1
1
  import { T as ThinkLevel } from './thinking-CBWSLel8.cjs';
2
- import { M as ModelConfig } from './llm-config-DeWNx1ig.cjs';
2
+ import { M as ModelConfig } from './llm-config-CEZvId7g.cjs';
3
3
 
4
4
  /**
5
5
  * Canonical Model Registry
package/dist/canonical-models.d.ts CHANGED
@@ -1,5 +1,5 @@
1
1
  import { T as ThinkLevel } from './thinking-CBWSLel8.js';
2
- import { M as ModelConfig } from './llm-config-CKOaFFdy.js';
2
+ import { M as ModelConfig } from './llm-config-Dxw7F1_U.js';
3
3
 
4
4
  /**
5
5
  * Canonical Model Registry
package/dist/canonical-models.js CHANGED
@@ -26,19 +26,56 @@ var ANTHROPIC_MODELS = {
26
26
  "claude-opus-4-6[1m]": { category: "reasoning", contextLength: 1e6, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 128e3 } },
27
27
  "claude-sonnet-4-6": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } },
28
28
  "claude-sonnet-4-6[1m]": { category: "chat", contextLength: 1e6, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } },
29
- "claude-haiku-4-5": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } }
29
+ "claude-haiku-4-5": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high", "xhigh", "max"], thinkingTokenLimit: { min: 1024, max: 64e3 } },
30
+ // Copilot-served variants (same model family on the Copilot API; context
31
+ // windows per the Copilot catalog where it diverges from the family default).
32
+ "claude-opus-4-5": { category: "reasoning", contextLength: 2e5, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true },
33
+ "claude-sonnet-4-5": { category: "chat", contextLength: 2e5, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true },
34
+ "claude-sonnet-4": { category: "chat", contextLength: 216e3, maxTokens: 64e3, vision: true, functionCall: true, reasoning: true }
30
35
  };
31
36
  var GEMINI_MODELS = {
32
37
  "gemini-3.5-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 24576 } },
33
38
  "gemini-3-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 24576 } },
34
- "gemini-3.1-pro": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 128, max: 32768 } }
39
+ "gemini-3.1-pro": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 128, max: 32768 } },
40
+ // Copilot-served variants (chat-completions wire on the Copilot API).
41
+ "gemini-3.1-pro-preview": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"] },
42
+ "gemini-3.6-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"] },
43
+ "gemini-3.7-flash": { category: "chat", contextLength: 1e6, maxTokens: 65536, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"] },
44
+ "gemini-3-pro-preview": { category: "chat", contextLength: 128e3, maxTokens: 65536, vision: true, functionCall: true, reasoning: true },
45
+ "gemini-3-flash-preview": { category: "chat", contextLength: 128e3, maxTokens: 65536, vision: true, functionCall: true, reasoning: true },
46
+ "gemini-2.5-pro": { category: "chat", contextLength: 128e3, maxTokens: 65536, vision: true, functionCall: true, reasoning: true }
35
47
  };
36
48
  var GROK_MODELS = {
37
49
  "grok-4.5": { category: "chat", contextLength: 5e5, maxTokens: 32768, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
38
50
  "grok-4.3": { category: "chat", contextLength: 1e6, maxTokens: 32768, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
39
51
  "grok-4.20": { category: "chat", contextLength: 131072, maxTokens: 32768, vision: true, functionCall: true },
40
52
  "grok-4.20-reasoning": { category: "reasoning", contextLength: 131072, maxTokens: 32768, reasoning: true, vision: true, functionCall: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
41
- "grok-4.20-multi-agent": { category: "reasoning", contextLength: 131072, maxTokens: 32768, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } }
53
+ "grok-4.20-multi-agent": { category: "reasoning", contextLength: 131072, maxTokens: 32768, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
54
+ // SuperGrok subscription ids (Responses wire). Upstream context/output caps
55
+ // per the audit source; the effort-taking models accept low/medium/high only
56
+ // (minimal→low, xhigh/max→high), and the omit-effort models carry
57
+ // `thinkingLevels: ['none']` so the effort negotiation drops the
58
+ // `reasoning.effort` field entirely (sending it upstream is a 400).
59
+ "grok-4.6": { category: "chat", contextLength: 5e5, maxTokens: 5e5, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
60
+ "grok-4.20-0309-non-reasoning": { category: "chat", contextLength: 2e6, maxTokens: 2e6, vision: true, functionCall: true, thinkingLevels: ["none"] },
61
+ "grok-4.20-0309-reasoning": { category: "reasoning", contextLength: 2e6, maxTokens: 2e6, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none"] },
62
+ "grok-4.20-multi-agent-0309": { category: "reasoning", contextLength: 2e6, maxTokens: 2e6, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
63
+ "grok-build": { category: "reasoning", contextLength: 512e3, maxTokens: 512e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none"] },
64
+ "grok-build-0.1": { category: "reasoning", contextLength: 256e3, maxTokens: 256e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none"] },
65
+ "grok-composer-2.5-fast": { category: "chat", contextLength: 2e5, maxTokens: 2e5, functionCall: true, thinkingLevels: ["none"] }
66
+ };
67
+ var COPILOT_MODELS = {
68
+ "gpt-5.1-codex-max": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
69
+ "gpt-5.1-codex-mini": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
70
+ "gpt-5.2-codex": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
71
+ "gpt-5.1-codex": { category: "code", contextLength: 272e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
72
+ "gpt-5.1": { category: "reasoning", contextLength: 264e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
73
+ "gpt-5-mini": { category: "reasoning", contextLength: 264e3, maxTokens: 128e3, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
74
+ "gpt-5.4-nano": { category: "reasoning", contextLength: 4e5, maxTokens: 128e3, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 128e3 } },
75
+ "grok-code-fast-1": { category: "code", contextLength: 192e3, maxTokens: 32768, functionCall: true, reasoning: true },
76
+ "raptor-mini": { category: "reasoning", contextLength: 4e5, maxTokens: 65536, functionCall: true, reasoning: true, vision: true },
77
+ "mai-code-1-flash-picker": { category: "code", contextLength: 256e3, maxTokens: 65536, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"] },
78
+ "mai-code-1.1-flash": { category: "code", contextLength: 256e3, maxTokens: 65536, functionCall: true, reasoning: true, vision: true, thinkingLevels: ["none", "low", "medium", "high"] }
42
79
  };
43
80
  var DEEPSEEK_MODELS = {
44
81
  "deepseek-v4-flash": { category: "chat", contextLength: 1e6, maxTokens: 384e3, functionCall: true, reasoning: true, thinkingLevels: ["none", "high", "max"], thinkingTokenLimit: { min: 0, max: 32768 } },
@@ -48,6 +85,8 @@ var DEEPSEEK_MODELS = {
48
85
  "deepseek-r1": { category: "reasoning", contextLength: 64e3, maxTokens: 8192, reasoning: true, thinkingTokenLimit: { min: 0, max: 32768 } }
49
86
  };
50
87
  var ZHIPU_MODELS = {
88
+ "glm-5.3": { category: "chat", contextLength: 1048576, maxTokens: 131072, functionCall: true },
89
+ "glm-5.3-flash": { category: "chat", contextLength: 1048576, maxTokens: 131072, functionCall: true },
51
90
  "glm-5.2": { category: "chat", contextLength: 1048576, maxTokens: 131072, functionCall: true },
52
91
  "glm-5": { category: "chat", contextLength: 2e5, maxTokens: 128e3, functionCall: true },
53
92
  "glm-5.1": { category: "chat", contextLength: 2e5, maxTokens: 128e3, functionCall: true },
@@ -82,10 +121,15 @@ var TENCENT_HUNYUAN_MODELS = {
82
121
  "hunyuan-turbos": { category: "chat", functionCall: true }
83
122
  };
84
123
  var MINIMAX_MODELS = {
124
+ "minimax-m3": { category: "code", contextLength: 204800, functionCall: true, reasoning: true },
125
+ "minimax-m2.7": { category: "code", contextLength: 204800, functionCall: true },
126
+ "minimax-m2.7-highspeed": { category: "code", contextLength: 204800, functionCall: true },
85
127
  "minimax-m2.5": { category: "code", contextLength: 204800, functionCall: true },
86
128
  "minimax-m2.5-highspeed": { category: "code", contextLength: 204800, functionCall: true },
129
+ "minimax-m2.5-lightning": { category: "code", contextLength: 204800, functionCall: true },
87
130
  "minimax-m2.1": { category: "code", contextLength: 204800, functionCall: true },
88
- "minimax-m2.1-highspeed": { category: "code", contextLength: 204800, functionCall: true }
131
+ "minimax-m2.1-lightning": { category: "code", contextLength: 204800, functionCall: true },
132
+ "minimax-m2": { category: "code", contextLength: 204800, functionCall: true }
89
133
  };
90
134
  var XIAOMI_MIMO_CANONICAL = {
91
135
  "mimo-v2.5": { category: "code", contextLength: 1e6, maxTokens: 32768, vision: true, functionCall: true, reasoning: true, thinkingLevels: ["none", "low", "medium", "high"], thinkingTokenLimit: { min: 0, max: 32768 } },
@@ -127,6 +171,7 @@ var VENDOR_GROUPS = [
127
171
  { name: "ANTHROPIC_MODELS", entries: ANTHROPIC_MODELS },
128
172
  { name: "GEMINI_MODELS", entries: GEMINI_MODELS },
129
173
  { name: "GROK_MODELS", entries: GROK_MODELS },
174
+ { name: "COPILOT_MODELS", entries: COPILOT_MODELS },
130
175
  { name: "DEEPSEEK_MODELS", entries: DEEPSEEK_MODELS },
131
176
  { name: "ZHIPU_MODELS", entries: ZHIPU_MODELS },
132
177
  { name: "KIMI_MODELS", entries: KIMI_MODELS },
package/dist/endpoint-resolver.d.cts CHANGED
@@ -1,4 +1,4 @@
1
- import { L as LLMProvider } from './llm-config-DeWNx1ig.cjs';
1
+ import { L as LLMProvider } from './llm-config-CEZvId7g.cjs';
2
2
  import './thinking-CBWSLel8.cjs';
3
3
 
4
4
  /**
package/dist/endpoint-resolver.d.ts CHANGED
@@ -1,4 +1,4 @@
1
- import { L as LLMProvider } from './llm-config-CKOaFFdy.js';
1
+ import { L as LLMProvider } from './llm-config-Dxw7F1_U.js';
2
2
  import './thinking-CBWSLel8.js';
3
3
 
4
4
  /**