yymaxapi 1.0.120 → 1.0.123

package/README.md

@@ -47,22 +47,22 @@ set OPENCLAW_CODEX_KEY=你的Key && npx yymaxapi@latest preset-codex
 **方式四：完全自动化（无交互）**
 ```bash
 # macOS/Linux
-OPENCLAW_CLAUDE_KEY="你的Key" npx yymaxapi@latest preset-claude --model claude-opus-4-7 --set-primary true --force --test true
+OPENCLAW_CLAUDE_KEY="你的Key" npx yymaxapi@latest preset-claude --model claude-opus-4-8 --set-primary true --force --test true
 
 # Windows PowerShell
-$env:OPENCLAW_CLAUDE_KEY="你的Key"; npx yymaxapi@latest preset-claude --model claude-opus-4-7 --set-primary true --force --test true
+$env:OPENCLAW_CLAUDE_KEY="你的Key"; npx yymaxapi@latest preset-claude --model claude-opus-4-8 --set-primary true --force --test true
 
 # Windows CMD
-set OPENCLAW_CLAUDE_KEY=你的Key && npx yymaxapi@latest preset-claude --model claude-opus-4-7 --set-primary true --force --test true
+set OPENCLAW_CLAUDE_KEY=你的Key && npx yymaxapi@latest preset-claude --model claude-opus-4-8 --set-primary true --force --test true
 
 # macOS/Linux
-OPENCLAW_CODEX_KEY="你的Key" npx yymaxapi@latest preset-codex --model gpt-5.4 --set-primary true --force --test true
+OPENCLAW_CODEX_KEY="你的Key" npx yymaxapi@latest preset-codex --model gpt-5.5 --set-primary true --force --test true
 
 # Windows PowerShell
-$env:OPENCLAW_CODEX_KEY="你的Key"; npx yymaxapi@latest preset-codex --model gpt-5.4 --set-primary true --force --test true
+$env:OPENCLAW_CODEX_KEY="你的Key"; npx yymaxapi@latest preset-codex --model gpt-5.5 --set-primary true --force --test true
 
 # Windows CMD
-set OPENCLAW_CODEX_KEY=你的Key && npx yymaxapi@latest preset-codex --model gpt-5.4 --set-primary true --force --test true
+set OPENCLAW_CODEX_KEY=你的Key && npx yymaxapi@latest preset-codex --model gpt-5.5 --set-primary true --force --test true
 ```
 
 ## 重要说明

package/bin/yymaxapi.js

@@ -99,13 +99,17 @@ const FALLBACK_ENDPOINTS = [
 
 const DEFAULT_CLAUDE_MODELS = [
   {
-    "id": "claude-sonnet-4-6",
-    "name": "Claude Sonnet 4.6"
+    "id": "claude-opus-4-8",
+    "name": "Claude Opus 4.8"
   },
   {
     "id": "claude-opus-4-7",
     "name": "Claude Opus 4.7"
   },
+  {
+    "id": "claude-sonnet-4-6",
+    "name": "Claude Sonnet 4.6"
+  },
   {
     "id": "claude-haiku-4-5",
     "name": "Claude Haiku 4.5"
@@ -113,6 +117,10 @@ const DEFAULT_CLAUDE_MODELS = [
 ];
 
 const DEFAULT_CODEX_MODELS = [
+  {
+    "id": "gpt-5.5",
+    "name": "GPT 5.5"
+  },
   {
     "id": "gpt-5.4",
     "name": "GPT 5.4"
@@ -127,8 +135,8 @@ const DEFAULT_API_CONFIG = {
   "claude": {
     "urlSuffix": "/claude",
     "api": "anthropic-messages",
-    "contextWindow": 200000,
-    "maxTokens": 8192,
+    "contextWindow": 1000000,
+    "maxTokens": 128000,
     "providerName": "yunyi-claude"
   },
   "codex": {
@@ -243,11 +251,11 @@ const CODEX_MODELS = PRESETS.models.codex;
 const API_CONFIG = PRESETS.apiConfig;
 
 function getDefaultClaudeModel() {
-  return CLAUDE_MODELS[0] || { id: 'claude-sonnet-4-6', name: 'Claude Sonnet 4.6' };
+  return CLAUDE_MODELS[0] || { id: 'claude-opus-4-8', name: 'Claude Opus 4.8' };
 }
 
 function getDefaultCodexModel() {
-  return CODEX_MODELS[0] || { id: 'gpt-5.4', name: 'GPT 5.4' };
+  return CODEX_MODELS[0] || { id: 'gpt-5.5', name: 'GPT 5.5' };
 }
 
 function getExternalClaudeProviderKey() {
@@ -337,6 +345,12 @@ function resolveHermesModelType(modelId, preferredType = '') {
   return requestedType || 'claude';
 }
 
+function getKnownModelName(modelId) {
+  const normalized = String(modelId || '').trim();
+  const model = [...CLAUDE_MODELS, ...CODEX_MODELS].find(item => item.id === normalized);
+  return model?.name || normalized || '未选择模型';
+}
+
 async function promptHermesModelSelection(args = {}, message = '选择 Hermes 默认模型:') {
   const requestedModel = (args['model-id'] || args.model || '').toString().trim();
   const requestedType = normalizeHermesModelType(
@@ -1131,7 +1145,7 @@ function writeClaudeCodeSettings(baseUrl, apiKey, modelId = getDefaultClaudeMode
     }
     const normalizedBaseUrl = baseUrl.replace(/\/+$/, '');
     const normalizedModelId = String(modelId || getDefaultClaudeModel().id).trim() || getDefaultClaudeModel().id;
-    const pinnedOpusModel = (CLAUDE_MODELS.find(model => /opus/i.test(model.id)) || {}).id || 'claude-opus-4-7';
+    const pinnedOpusModel = (CLAUDE_MODELS.find(model => /opus/i.test(model.id)) || {}).id || 'claude-opus-4-8';
     const pinnedSonnetModel = (CLAUDE_MODELS.find(model => /sonnet/i.test(model.id)) || {}).id || 'claude-sonnet-4-6';
     const pinnedHaikuModel = (CLAUDE_MODELS.find(model => /haiku/i.test(model.id)) || {}).id || 'claude-haiku-4-5';
     settings.apiBaseUrl = normalizedBaseUrl;
@@ -1150,7 +1164,7 @@ function writeClaudeCodeSettings(baseUrl, apiKey, modelId = getDefaultClaudeMode
   } catch { /* 非关键，静默失败 */ }
 }
 
-function writeCodexConfig(baseUrl, apiKey, modelId = 'gpt-5.4') {
+function writeCodexConfig(baseUrl, apiKey, modelId = 'gpt-5.5') {
   const codexDir = path.join(os.homedir(), '.codex');
   if (!fs.existsSync(codexDir)) fs.mkdirSync(codexDir, { recursive: true });
 
@@ -1185,6 +1199,7 @@ function writeCodexConfig(baseUrl, apiKey, modelId = 'gpt-5.4') {
       }
       existing = existing.replace(/\[model_providers\.openclaw-relay\]\n(?:(?!\[)[^\n]*\n?)*/g, '');
       existing = existing.replace(/# >>> yunyi opencode >>>[\s\S]*?# <<< yunyi opencode <<<\n?/g, '');
+      existing = existing.replace(/# >>> maxapi opencode >>>[\s\S]*?# <<< maxapi opencode <<<\n?/g, '');
       existing = existing.replace(/\n{3,}/g, '\n\n').trim();
     }
     const hasExistingTopLevelModelConfig = /^model\s*=\s*"[^"]*"\s*$/m.test(existing)
@@ -1303,19 +1318,54 @@ function writeOpencodeConfig(claudeBaseUrl, codexBaseUrl, apiKey, defaultModelKe
       content = fs.readFileSync(codexConfigPath, 'utf8');
     }
 
-    // 移除旧标记块（仅移除 yunyi opencode 自己的）
+    // 移除旧标记块（仅移除 opencode 自己写入的兼容 provider）
     content = content.replace(/# >>> yunyi opencode >>>[\s\S]*?# <<< yunyi opencode <<<\n?/g, '');
-
-    // 移除旧版 yunyi-cli 写入的 [model_providers.yunyi] 和 [model_providers.yunyi-claude]（不动 yunyi-codex，由 writeCodexConfig 管理）
-    content = content.replace(/\[model_providers\.yunyi\]\n(?:(?!\[)[^\n]*\n?)*/g, '');
-    content = content.replace(/\[model_providers\.yunyi-claude\]\n(?:(?!\[)[^\n]*\n?)*/g, '');
+    content = content.replace(/# >>> maxapi opencode >>>[\s\S]*?# <<< maxapi opencode <<<\n?/g, '');
+
+    // 移除旧版 yunyi-cli / maxai-cli 写入的同名 provider，随后写入当前受管块
+    const removableProviderKeys = [
+      'yunyi',
+      'yunyi-claude',
+      'yunyi-codex',
+      'maxapi',
+      'maxapi-codex',
+      claudeProviderKey,
+      codexProviderKey
+    ];
+    for (const providerKey of [...new Set(removableProviderKeys)]) {
+      const escapedProviderKey = escapeRegExp(providerKey);
+      content = content.replace(new RegExp(`\\[model_providers\\.${escapedProviderKey}\\]\\n(?:(?!\\[)[^\\n]*\\n?)*`, 'g'), '');
+    }
 
     // 清理多余空行
     content = content.replace(/\n{3,}/g, '\n\n').trim();
 
-    // yunyi-codex provider 由 writeCodexConfig 统一管理，此处不再重复写入
+    const markerPrefix = claudeProviderKey.startsWith('maxapi') || codexProviderKey.startsWith('maxapi') ? 'maxapi' : 'yunyi';
+    const marker = `# >>> ${markerPrefix} opencode >>>`;
+    const markerEnd = `# <<< ${markerPrefix} opencode <<<`;
+    let normalizedCodexTomlUrl = codexUrl;
+    if (normalizedCodexTomlUrl && !normalizedCodexTomlUrl.endsWith('/v1')) normalizedCodexTomlUrl += '/v1';
+
+    const opencodeProviderBlock = [
+      marker,
+      `[model_providers.${claudeProviderKey}]`,
+      `name = "${brandPrefix} Claude"`,
+      `base_url = "${claudeUrl}/v1"`,
+      `wire_api = "anthropic"`,
+      `experimental_bearer_token = "${apiKey}"`,
+      ...(normalizedCodexTomlUrl ? [
+        '',
+        `[model_providers.${codexProviderKey}]`,
+        `name = "${brandPrefix} Codex"`,
+        `base_url = "${normalizedCodexTomlUrl}"`,
+        `wire_api = "responses"`,
+        `experimental_bearer_token = "${apiKey}"`
+      ] : []),
+      markerEnd
+    ].join('\n');
 
-    fs.writeFileSync(codexConfigPath, content, 'utf8');
+    const nextContent = [content, opencodeProviderBlock].filter(Boolean).join('\n\n') + '\n';
+    fs.writeFileSync(codexConfigPath, nextContent, 'utf8');
 
     // ---- 3. ~/.codex/auth.json ----
     const authPath = path.join(codexDir, 'auth.json');
@@ -2319,15 +2369,15 @@ const YYMAXAPI_OPENCLAW_MAIN_AGENT_ID = 'main';
 const YYMAXAPI_OPENCLAW_ALT_CLAUDE_AGENT_ID = 'yunyi-claude';
 const YYMAXAPI_OPENCLAW_GPT_AGENT_ID = 'yunyi-gpt';
 const YYMAXAPI_OPENCLAW_LEGACY_GPT_AGENT_IDS = ['gpt'];
-const YYMAXAPI_OPENCLAW_CLAUDE_PRIMARY = `${YYMAXAPI_OPENCLAW_CLAUDE_PROVIDER}/claude-sonnet-4-6`;
+const YYMAXAPI_OPENCLAW_CLAUDE_PRIMARY = `${YYMAXAPI_OPENCLAW_CLAUDE_PROVIDER}/claude-opus-4-8`;
 const YYMAXAPI_OPENCLAW_CLAUDE_FALLBACK = `${YYMAXAPI_OPENCLAW_CLAUDE_PROVIDER}/claude-opus-4-7`;
-const YYMAXAPI_OPENCLAW_GPT_PRIMARY = `${YYMAXAPI_OPENCLAW_GPT_PROVIDER}/gpt-5.4`;
-const YYMAXAPI_OPENCLAW_GPT_FALLBACK = `${YYMAXAPI_OPENCLAW_GPT_PROVIDER}/gpt-5.3-codex`;
+const YYMAXAPI_OPENCLAW_GPT_PRIMARY = `${YYMAXAPI_OPENCLAW_GPT_PROVIDER}/gpt-5.5`;
+const YYMAXAPI_OPENCLAW_GPT_FALLBACK = `${YYMAXAPI_OPENCLAW_GPT_PROVIDER}/gpt-5.4`;
 const YYMAXAPI_MANAGED_MAIN_NAMES = new Set(['Claude', 'claude', 'yunyi-claude', 'claude-yunyi']);
 const YYMAXAPI_MANAGED_GPT_NAMES = new Set(['GPT', 'Codex', 'gpt', 'yunyi-gpt', 'yunyi-codex']);
 const YYMAXAPI_MANAGED_MULTIMODAL_MODELS = {
-  claude: new Set(['claude-sonnet-4-6', 'claude-opus-4-7', 'claude-haiku-4-5']),
-  codex: new Set(['gpt-5.4', 'gpt-5.3-codex'])
+  claude: new Set(['claude-opus-4-8', 'claude-opus-4-7', 'claude-sonnet-4-6', 'claude-haiku-4-5']),
+  codex: new Set(['gpt-5.5', 'gpt-5.4', 'gpt-5.3-codex'])
 };
 
 function isYymaxapiOpenClawBuild() {
@@ -2380,7 +2430,7 @@ function getManagedYunyiModelInput(type, modelId) {
 
 function getManagedYunyiProviderModels(type) {
   const apiConfig = type === 'claude'
-    ? { contextWindow: API_CONFIG.claude?.contextWindow || 200000, maxTokens: API_CONFIG.claude?.maxTokens || 8192 }
+    ? { contextWindow: API_CONFIG.claude?.contextWindow || 1000000, maxTokens: API_CONFIG.claude?.maxTokens || 128000 }
     : { contextWindow: API_CONFIG.codex?.contextWindow || 1000000, maxTokens: API_CONFIG.codex?.maxTokens || 128000 };
 
   return getManagedYunyiModelCatalog(type).map(model => ({
@@ -2449,7 +2499,7 @@ function normalizeManagedYunyiModelState(type, modelState = {}, options = {}) {
   const canPreserve = !options.force && currentPrimary && knownKeys.includes(currentPrimary);
   const primary = canPreserve ? currentPrimary : fallbackPrimary;
 
-  let fallbacks = Array.isArray(modelState.fallbacks) ? modelState.fallbacks : [];
+  let fallbacks = !options.force && Array.isArray(modelState.fallbacks) ? modelState.fallbacks : [];
   fallbacks = [...new Set(
     fallbacks
       .map(item => canonicalizeManagedYunyiModelKey(item, type))
@@ -2470,7 +2520,7 @@ function normalizeManagedYunyiDefaultsModel(modelState = {}, options = {}) {
     ? canonicalizeManagedYunyiModelKey(currentPrimary)
     : YYMAXAPI_OPENCLAW_CLAUDE_PRIMARY;
 
-  let fallbacks = Array.isArray(modelState.fallbacks) ? modelState.fallbacks : [];
+  let fallbacks = !options.force && Array.isArray(modelState.fallbacks) ? modelState.fallbacks : [];
   fallbacks = [...new Set(
     fallbacks
       .map(item => {
@@ -2962,7 +3012,7 @@ function normalizeManagedYunyiImageModel(modelState = {}, options = {}) {
     ? canonicalizeManagedYunyiModelKey(currentPrimary)
     : YYMAXAPI_OPENCLAW_CLAUDE_PRIMARY;
 
-  let fallbacks = Array.isArray(modelState.fallbacks) ? modelState.fallbacks : [];
+  let fallbacks = !options.force && Array.isArray(modelState.fallbacks) ? modelState.fallbacks : [];
   fallbacks = [...new Set(
     fallbacks
       .map(item => {
@@ -3071,7 +3121,16 @@ function applyManagedYunyiOpenClawLayout(config, options = {}) {
   }
 
   if (shouldManageYunyiImageDefaults(config)) {
-    const currentImageModel = normalizeDefaultModelSelection(config.agents.defaults.imageModel, config);
+    const hasExplicitImageModel = config.agents.defaults.imageModel
+      && typeof config.agents.defaults.imageModel === 'object'
+      && !Array.isArray(config.agents.defaults.imageModel)
+      && (
+        typeof config.agents.defaults.imageModel.primary === 'string'
+        || Array.isArray(config.agents.defaults.imageModel.fallbacks)
+      );
+    const currentImageModel = hasExplicitImageModel
+      ? normalizeDefaultModelSelection(config.agents.defaults.imageModel, config)
+      : {};
     const nextImageModel = normalizeManagedYunyiImageModel(currentImageModel, options);
     if (JSON.stringify(currentImageModel) !== JSON.stringify(nextImageModel)) {
       config.agents.defaults.imageModel = nextImageModel;
@@ -5591,10 +5650,10 @@ async function autoActivate(paths, args = {}) {
   const isClaudePrimary = selectedType === 'claude';
   const claudeModelId = isClaudePrimary
     ? selectedModelId
-    : CLAUDE_MODELS[0]?.id || 'claude-sonnet-4-6';
+    : CLAUDE_MODELS[0]?.id || 'claude-opus-4-8';
   const codexModelId = !isClaudePrimary
     ? selectedModelId
-    : CODEX_MODELS[0]?.id || 'gpt-5.4';
+    : CODEX_MODELS[0]?.id || 'gpt-5.5';
 
   const claudeModel = CLAUDE_MODELS.find(m => m.id === claudeModelId) || { id: claudeModelId, name: claudeModelId };
   const codexModel = CODEX_MODELS.find(m => m.id === codexModelId) || { id: codexModelId, name: codexModelId };
@@ -5957,10 +6016,11 @@ async function activateHermes(paths, args = {}) {
 async function main() {
   console.clear();
 
-  console.log(chalk.cyan.bold('\n🔧 OpenClaw API 配置工具\n'));
+  console.log(chalk.cyan.bold('\n🔧 yymaxapi 配置工具\n'));
 
   const paths = getConfigPath();
-  console.log(chalk.gray(`配置文件: ${paths.openclawConfig}\n`));
+  console.log(chalk.gray(`OpenClaw 配置: ${paths.openclawConfig}`));
+  console.log(chalk.gray(`Hermes 配置: ${getHermesConfigPath()}\n`));
 
   // 首次运行备份
   if (backupOriginalConfig(paths.openclawConfig, paths.configDir)) {
@@ -5999,10 +6059,15 @@ async function main() {
 
   while (true) {
     // 显示当前状态
-    const statusLine = getConfigStatusLine(paths);
-    if (statusLine) {
+    const statusLines = [
+      getConfigStatusLine(paths),
+      getHermesStatusLine()
+    ].filter(Boolean);
+    if (statusLines.length > 0) {
       console.log(chalk.gray('─'.repeat(40)));
-      console.log(statusLine);
+      for (const statusLine of statusLines) {
+        console.log(statusLine);
+      }
       console.log(chalk.gray('─'.repeat(40)) + '\n');
     }
 
@@ -6125,7 +6190,7 @@ function getConfigStatusLine(paths) {
     }
 
     if (parts.length === 0 && otherCount === 0) {
-      return chalk.gray('当前状态: 未配置任何模型');
+      return chalk.gray('OpenClaw: 未配置任何模型');
     }
 
     const primaryModelId = primary.split('/')[1] || '';
@@ -6133,7 +6198,39 @@ function getConfigStatusLine(paths) {
 
     const gwEnv = detectGatewayEnv();
     const envTag = gwEnv === 'wsl' ? chalk.cyan(' [WSL]') : '';
-    return chalk.gray('当前状态: ') + parts.join('  ') + chalk.gray('  (✓ 主模型  ○ 已配置)') + modelTag + envTag;
+    return chalk.gray('OpenClaw: ') + parts.join('  ') + chalk.gray('  (✓ 主模型  ○ 已配置)') + modelTag + envTag;
+  } catch {
+    return null;
+  }
+}
+
+function getHermesStatusLine() {
+  try {
+    let config = readHermesCliConfig();
+    let wslMirror = null;
+
+    if (process.platform === 'win32') {
+      wslMirror = getHermesWslMirrorInfo();
+      if (!config.configured && (wslMirror.configPath || wslMirror.envPath)) {
+        config = readHermesWslCliConfig(wslMirror);
+      }
+    }
+
+    if (!config.configured) return null;
+
+    const modelName = getKnownModelName(config.modelId);
+    const readyMark = config.apiKey ? chalk.green('✓') : chalk.yellow('○');
+    const modelTag = config.modelId ? chalk.cyan(` [模型: ${config.modelId}]`) : '';
+    const protocolTag = config.type === 'codex' ? chalk.cyan(' [GPT]') : chalk.cyan(' [Claude]');
+    const hasWslTarget = !!(wslMirror?.sourceConfigPath || wslMirror?.commandPath);
+    let envTag = '';
+    if (hasWslTarget && wslMirror?.target?.user) {
+      envTag = chalk.cyan(` [WSL ${wslMirror.target.user}]`);
+    } else if (hasWslTarget) {
+      envTag = chalk.cyan(' [WSL]');
+    }
+
+    return chalk.gray('Hermes-Agent: ') + chalk.green(`${modelName} ${readyMark}`) + protocolTag + modelTag + envTag;
   } catch {
     return null;
   }
@@ -7414,7 +7511,7 @@ function testClaudeApi(baseUrl, apiKey, model) {
     const protocol = urlObj.protocol === 'https:' ? https : http;
 
     const postData = JSON.stringify({
-      model: model || 'claude-sonnet-4-6',
+      model: model || 'claude-opus-4-8',
       max_tokens: 150,
       messages: [{ role: 'user', content: '你是哪个模型？请用一句话回答你的模型名称和版本。' }]
     });
@@ -7474,7 +7571,7 @@ function testCodexApi(baseUrl, apiKey, model) {
     const protocol = urlObj.protocol === 'https:' ? https : http;
 
     const postData = JSON.stringify({
-      model: model || 'gpt-5.4',
+      model: model || 'gpt-5.5',
       input: '你是哪个模型？请用一句话回答你的模型名称和版本。'
     });
 

package/config/API/350/212/202/347/202/271/350/256/276/347/275/256.md

@@ -14,7 +14,8 @@
 {
   "endpoints": [
     { "name": "默认主节点", "url": "https://yunyi.yun" },
-    { "name": "CF国外节点1", "url": "https://cdn2.yunyi.yun" }
+    { "name": "备用 CDN 域名 1", "url": "https://cdn1.yunyi.yun" },
+    { "name": "备用 CDN 域名 2", "url": "https://cdn2.yunyi.yun" }
   ],
   "fallbackEndpoints": [
     { "name": "备用节点1", "url": "http://47.99.42.193" },
@@ -22,11 +23,13 @@
   ],
   "models": {
     "claude": [
-      { "id": "claude-sonnet-4-6", "name": "Claude Sonnet 4.6" },
+      { "id": "claude-opus-4-8", "name": "Claude Opus 4.8" },
       { "id": "claude-opus-4-7", "name": "Claude Opus 4.7" },
+      { "id": "claude-sonnet-4-6", "name": "Claude Sonnet 4.6" },
       { "id": "claude-haiku-4-5", "name": "Claude Haiku 4.5" }
     ],
     "codex": [
+      { "id": "gpt-5.5", "name": "GPT 5.5" },
       { "id": "gpt-5.4", "name": "GPT 5.4" },
       { "id": "gpt-5.3-codex", "name": "GPT 5.3 Codex" }
     ]
@@ -35,8 +38,8 @@
     "claude": {
       "urlSuffix": "/claude",
       "api": "anthropic-messages",
-      "contextWindow": 200000,
-      "maxTokens": 8192,
+      "contextWindow": 1000000,
+      "maxTokens": 128000,
       "providerName": "yunyi-claude"
     },
     "codex": {
@@ -70,7 +73,7 @@ npx yymaxapi@latest
 
 在腾讯云 OpenClaw 部署中，使用「自定义模型」接入云翼中转，已验证可用的配置：
 
-### Claude Sonnet 4.6
+### Claude Opus 4.8
 
 ```json
 {
@@ -79,13 +82,13 @@ npx yymaxapi@latest
   "api": "anthropic-messages",
   "api_key": "<你的云翼 API Key>",
   "model": {
-    "id": "claude-sonnet-4-6",
-    "name": "Claude Sonnet 4.6"
+    "id": "claude-opus-4-8",
+    "name": "Claude Opus 4.8"
   }
 }
 ```
 
-### GPT 5.4
+### GPT 5.5
 
 ```json
 {
@@ -94,8 +97,8 @@ npx yymaxapi@latest
   "api": "openai-completions",
   "api_key": "<你的云翼 API Key>",
   "model": {
-    "id": "gpt-5.4",
-    "name": "GPT 5.4"
+    "id": "gpt-5.5",
+    "name": "GPT 5.5"
   }
 }
 ```
@@ -103,8 +106,11 @@ npx yymaxapi@latest
 **注意事项：**
 - `base_url` 不要加 `/v1`，平台会自动拼接路径
 - 如需临时切到其他云翼域名，可直接把域名部分替换掉；`yymaxapi` 也支持在预设/一键激活命令中传 `--endpoint-url https://你的域名` 或 `--base-url https://你的域名/claude`
-- Claude 可用模型：`claude-sonnet-4-6`、`claude-opus-4-7`、`claude-haiku-4-5`
-- GPT 可用模型：`gpt-5.4`
+- Claude 可用模型：`claude-opus-4-8`、`claude-opus-4-7`、`claude-sonnet-4-6`、`claude-haiku-4-5`
+- Claude 默认模型：`claude-opus-4-8`
+- `claude-opus-4-8` 按当前官方规格声明为 1M context / 128K max output
+- GPT 可用模型：`gpt-5.5`、`gpt-5.4`、`gpt-5.3-codex`
+- GPT 默认模型：`gpt-5.5`
 - 已验证环境：腾讯云 OpenCloudOS，OpenClaw `2026.2.3-1`
 - 参考文档：https://cloud.tencent.com/developer/article/2624003
 
@@ -116,7 +122,7 @@ npx yymaxapi@latest
 
 在腾讯云 OpenClaw 的「自定义模型」里添加**两条**，`api_key` 用同一云翼 Key 即可。
 
-第一条 — Claude Sonnet 4.6：
+第一条 — Claude Opus 4.8：
 
 ```json
 {
@@ -125,13 +131,13 @@ npx yymaxapi@latest
   "api": "anthropic-messages",
   "api_key": "<你的云翼 API Key>",
   "model": {
-    "id": "claude-sonnet-4-6",
-    "name": "Claude Sonnet 4.6"
+    "id": "claude-opus-4-8",
+    "name": "Claude Opus 4.8"
   }
 }
 ```
 
-第二条 — GPT 5.4：
+第二条 — GPT 5.5：
 
 ```json
 {
@@ -140,8 +146,8 @@ npx yymaxapi@latest
   "api": "openai-completions",
   "api_key": "<你的云翼 API Key>",
   "model": {
-    "id": "gpt-5.4",
-    "name": "GPT 5.4"
+    "id": "gpt-5.5",
+    "name": "GPT 5.5"
   }
 }
 ```

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "yymaxapi",
-  "version": "1.0.120",
+  "version": "1.0.123",
   "description": "跨平台 OpenClaw/Clawdbot 配置管理工具 - 管理中转地址、模型切换、API Keys、测速优化",
   "main": "bin/yymaxapi.js",
   "bin": {