loom-agent 1.3.3 → 1.5.0

package/src/providers/openai-compat.js

@@ -52,7 +52,7 @@ function formatTools(tools) {
   }));
 }
 
-function buildRequest(messages, options) {
+function buildRequest(messages, options, providerId) {
   const body = {
     model: options.model,
     messages: formatMessages(messages, options),
@@ -60,6 +60,12 @@ function buildRequest(messages, options) {
     max_tokens: options.maxTokens || 8192,
     temperature: options.temperature ?? 0.7,
   };
+  // OpenRouter/TokenRouter prompt caching: top-level "cache_control" true
+  // (auto-detects prefix-cache-capable models, ignores it otherwise). Cuts
+  // input cost on cache hits (~10% of full price) after the first turn.
+  if (providerId === 'openrouter' || providerId === 'tokenrouter') {
+    body.cache_control = true;
+  }
   // Reasoning models (o1/o3/o4/gpt-5, tagged 'reasoning' by the session):
   // push the effort to "high" so the model reasons deeply on every step —
   // opencode-style multi-pass thinking. Other reasoning families (DeepSeek
@@ -110,27 +116,43 @@ function sanitizeErrBody(msg) {
 function wrapErr(err, modelId, envKeyHint, baseURL) {
   const raw = err.message || 'Unknown error';
   const msg = sanitizeErrBody(raw);
-  if (!err.status) return new Error(`${envKeyHint || 'API'} error: ${msg}`);
-  const status = Number(err.status);
-  const hint = String(envKeyHint || 'api').toLowerCase();
-  if (status === 401) {
-    return new Error(`${envKeyHint || 'API'} 401 Unauthorized: the API key is invalid or expired. Run /connect ${hint} and paste a new key.`);
-  }
-  if (status === 403) {
-    return new Error(`${envKeyHint || 'API'} 403 Forbidden: the API key is not authorized for ${modelId || 'this model'}. You may need to accept the model's terms on the provider site or use a key with access. Run /connect ${hint} to change the key.`);
-  }
-  if (status === 402) return new Error(`${envKeyHint || 'API'} quota exceeded (${err.status}). The model may not be available on your billing tier for model ${(modelId || '?')}.`);
-  if (status === 404) {
-    const modelHint = modelId ? ` ${modelId}` : '';
-    return new Error(`${envKeyHint || 'API'} 404: model${modelHint} does not exist (or was removed) on this provider. Try /models and pick a listed model, or /model ${modelId ? 'to switch' : ''}.`);
-  }
-  if (err.error && err.error.message) return new Error(`${envKeyHint || 'API'} error ${status}: ${sanitizeErrBody(err.error.message)}`);
-  if (err.body && err.body.message) return new Error(`${envKeyHint || 'API'} error ${status}: ${sanitizeErrBody(err.body.message)}`);
-  if (!msg || msg.indexOf('no body') >= 0) {
-    const modelHint = modelId ? ` (model: ${modelId})` : '';
-    return new Error(`${envKeyHint || 'API'} error ${status}${modelHint}. Check model name or URL:\n${baseURL || '(default)'}`);
+  // Carry the HTTP status + Retry-After onto the wrapped error so the
+  // session loop can tell a rate-limit (wait+retry) from a quota switch.
+  const out = /** @type {any} */ (!err.status ? new Error(`${envKeyHint || 'API'} error: ${msg}`) : (() => {
+    const status = Number(err.status);
+    const hint = String(envKeyHint || 'api').toLowerCase();
+    if (status === 401) {
+      return new Error(`${envKeyHint || 'API'} 401 Unauthorized: the API key is invalid or expired. Run /connect ${hint} and paste a new key.`);
+    }
+    if (status === 403) {
+      return new Error(`${envKeyHint || 'API'} 403 Forbidden: the API key is not authorized for ${modelId || 'this model'}. You may need to accept the model's terms on the provider site or use a key with access. Run /connect ${hint} to change the key.`);
+    }
+    if (status === 402) return new Error(`${envKeyHint || 'API'} quota exceeded (${err.status}). The model may not be available on your billing tier for model ${(modelId || '?')}.`);
+    if (status === 404) {
+      const modelHint = modelId ? ` ${modelId}` : '';
+      return new Error(`${envKeyHint || 'API'} 404: model${modelHint} does not exist (or was removed) on this provider. Try /models and pick a listed model, or /model ${modelId ? 'to switch' : ''}.`);
+    }
+    if (err.error && err.error.message) return new Error(`${envKeyHint || 'API'} error ${status}: ${sanitizeErrBody(err.error.message)}`);
+    if (err.body && err.body.message) return new Error(`${envKeyHint || 'API'} error ${status}: ${sanitizeErrBody(err.body.message)}`);
+    if (!msg || msg.indexOf('no body') >= 0) {
+      const modelHint = modelId ? ` (model: ${modelId})` : '';
+      return new Error(`${envKeyHint || 'API'} error ${status}${modelHint}. Check model name or URL:\n${baseURL || '(default)'}`);
+    }
+    return new Error(`${envKeyHint || 'API'} error ${status}: ${msg}`);
+  })());
+  out.status = err.status ? Number(err.status) : undefined;
+  // Retry-After from the openai SDK error headers (seconds or HTTP-date).
+  const hdrs = err.headers;
+  const ra = hdrs && (typeof hdrs.get === 'function' ? hdrs.get('retry-after') : hdrs['retry-after']);
+  if (ra) {
+    const secs = Number(ra);
+    if (Number.isFinite(secs) && secs >= 0) out.retryAfterMs = Math.min(secs * 1000, 60000);
+    else {
+      const t = Date.parse(ra);
+      if (Number.isFinite(t)) out.retryAfterMs = Math.min(Math.max(0, t - Date.now()), 60000);
+    }
   }
-  return new Error(`${envKeyHint || 'API'} error ${status}: ${msg}`);
+  return out;
 }
 
 async function retryWithBackoff(fn, maxTries, envKeyHint) {
@@ -142,10 +164,20 @@ async function retryWithBackoff(fn, maxTries, envKeyHint) {
       lastErr = err;
       if (isAbortError(err)) throw err;
       const status = err.status || err.code;
-      if (status !== 429 && status !== 503 && status !== 502) throw lastErr;
-      if (attempt > 0) {
-        const delay = Math.min(1000 * Math.pow(3, attempt - 1), 30000);
-        await new Promise((resolve) => setTimeout(resolve, delay));
+      if (status !== 429 && status !== 503 && status !== 502 && status !== 529) throw lastErr;
+      {
+        // Honor Retry-After when the provider sends one (even on the first
+        // retry); otherwise the first retry is immediate (transient blip)
+        // and later ones back off exponentially. Jitter avoids a thundering
+        // herd of parallel retries.
+        let delay = attempt > 0 ? Math.min(1000 * Math.pow(3, attempt - 1), 30000) : 0;
+        const hdrs = err.headers;
+        const ra = hdrs && (typeof hdrs.get === 'function' ? hdrs.get('retry-after') : hdrs['retry-after']);
+        if (ra) {
+          const secs = Number(ra);
+          if (Number.isFinite(secs) && secs >= 0) delay = Math.min(secs * 1000, 60000);
+        }
+        await new Promise((resolve) => setTimeout(resolve, delay + (delay > 0 ? Math.floor(Math.random() * 250) : 0)));
       }
     }
   }
@@ -194,7 +226,7 @@ function createOpenAICompatProvider({ getKey, providerId, envKeyHint, clientFact
 
   async function chat(messages, options = {}) {
     const client = getClient();
-    const req = buildRequest(messages, options);
+    const req = buildRequest(messages, options, providerId);
     const requestOpts = { signal: options.signal };
     try {
       const resp = await retryWithBackoff(() => client.chat.completions.create(req, requestOpts), 4, envKeyHint);
@@ -207,7 +239,7 @@ function createOpenAICompatProvider({ getKey, providerId, envKeyHint, clientFact
 
   async function stream(messages, options = {}, onDelta, onReasoning) {
     const client = getClient();
-    const req = { ...buildRequest(messages, options), stream: true, stream_options: { include_usage: true } };
+    const req = { ...buildRequest(messages, options, providerId), stream: true, stream_options: { include_usage: true } };
     const requestOpts = { signal: options.signal };
     try {
       const s = await retryWithBackoff(() => client.chat.completions.create(req, requestOpts), 4, envKeyHint);

package/src/providers/openai.js

@@ -11,24 +11,48 @@ const provider = {
     envKeyHint: 'OPENAI',
   }),
   models: [
-  // Flagship / GPT-5 Series
-  { id: 'gpt-5', name: 'GPT-5', provider: 'openai', tags: ['frontier'], context: 400000, priceIn: 1.25, priceOut: 10 },
-  { id: 'gpt-5-mini', name: 'GPT-5 Mini', provider: 'openai', tags: ['fast'], context: 400000, priceIn: 0.25, priceOut: 2.00 },
-  { id: 'gpt-5-nano', name: 'GPT-5 Nano', provider: 'openai', tags: ['cheap', 'fast'], context: 400000, priceIn: 0.10, priceOut: 0.40 },
-  { id: 'gpt-5-chat-latest', name: 'GPT-5 Chat (Latest)', provider: 'openai', tags: ['frontier'], context: 400000, priceIn: 5.00, priceOut: 30.00 },
-
-  // GPT-4 Series & Omni Models
-  { id: 'gpt-4o', name: 'GPT-4o', provider: 'openai', tags: ['legacy', 'multimodal'], context: 128000, priceIn: 2.50, priceOut: 10.00 },
-  { id: 'gpt-4o-mini', name: 'GPT-4o Mini', provider: 'openai', tags: ['fast', 'multimodal'], context: 128000, priceIn: 0.15, priceOut: 0.60 },
-  { id: 'gpt-4.1', name: 'GPT-4.1', provider: 'openai', tags: ['frontier'], context: 1048576, priceIn: 2.00, priceOut: 8.00 },
-
-  // O-Series (Reasoning) Models
-  { id: 'o3', name: 'o3', provider: 'openai', tags: ['reasoning', 'frontier'], context: 200000, priceIn: 2.00, priceOut: 8.00 },
-  { id: 'o3-mini', name: 'o3 Mini', provider: 'openai', tags: ['reasoning', 'fast'], context: 200000, priceIn: 1.10, priceOut: 4.40 },
-  { id: 'o4-mini', name: 'o4 Mini', provider: 'openai', tags: ['reasoning', 'fast'], context: 200000, priceIn: 1.10, priceOut: 4.40 },
-  { id: 'o1', name: 'o1', provider: 'openai', tags: ['reasoning'], context: 200000, priceIn: 15.00, priceOut: 60.00 },
-  { id: 'o1-pro', name: 'o1 Pro', provider: 'openai', tags: ['reasoning', 'pro'], context: 200000, priceIn: 150.00, priceOut: 600.00 },
-  { id: 'o1-mini', name: 'o1 Mini', provider: 'openai', tags: ['reasoning', 'cheap'], context: 128000, priceIn: 1.10, priceOut: 4.40 },
+  // ── GPT-6 Series ──
+  { id: 'gpt-6-astra', name: 'GPT-6 Astra', provider: 'openai', tags: ['frontier', 'reasoning'], context: 1050000, priceIn: 10, priceOut: 50 },
+  { id: 'gpt-6-sol', name: 'GPT-6 Sol', provider: 'openai', tags: ['reasoning'], context: 1050000, priceIn: 2, priceOut: 10 },
+  { id: 'gpt-6.1-sol', name: 'GPT-6.1 Sol', provider: 'openai', tags: ['reasoning'], context: 1050000, priceIn: 2, priceOut: 10 },
+  { id: 'gpt-6-luna', name: 'GPT-6 Luna', provider: 'openai', tags: ['fast', 'cheap', 'reasoning'], context: 1050000, priceIn: 0.1, priceOut: 0.5 },
+
+  // ── GPT-5.6 / 5.5 / 5.4 ──
+  { id: 'gpt-5.6', name: 'GPT-5.6', provider: 'openai', tags: ['frontier', 'reasoning'], context: 1050000, priceIn: 4, priceOut: 20 },
+  { id: 'gpt-5.6-sol', name: 'GPT-5.6 Sol', provider: 'openai', tags: ['reasoning'], context: 1050000, priceIn: 4, priceOut: 20 },
+  { id: 'gpt-5.6-terra', name: 'GPT-5.6 Terra', provider: 'openai', tags: ['reasoning'], context: 1050000, priceIn: 2, priceOut: 12 },
+  { id: 'gpt-5.6-luna', name: 'GPT-5.6 Luna', provider: 'openai', tags: ['fast', 'cheap', 'reasoning'], context: 1050000, priceIn: 0.2, priceOut: 1.2 },
+  { id: 'gpt-5.5', name: 'GPT-5.5', provider: 'openai', tags: ['frontier', 'reasoning'], context: 1050000, priceIn: 5, priceOut: 30 },
+  { id: 'gpt-5.5-pro', name: 'GPT-5.5 Pro', provider: 'openai', tags: ['reasoning', 'pro'], context: 1050000, priceIn: 30, priceOut: 180 },
+  { id: 'gpt-5.4', name: 'GPT-5.4', provider: 'openai', tags: ['reasoning'], context: 1050000, priceIn: 2.5, priceOut: 15 },
+  { id: 'gpt-5.4-mini', name: 'GPT-5.4 Mini', provider: 'openai', tags: ['fast', 'reasoning'], context: 400000, priceIn: 0.75, priceOut: 4.5 },
+  { id: 'gpt-5.4-nano', name: 'GPT-5.4 Nano', provider: 'openai', tags: ['cheap', 'fast', 'reasoning'], context: 400000, priceIn: 0.2, priceOut: 1.25 },
+  { id: 'gpt-5.4-pro', name: 'GPT-5.4 Pro', provider: 'openai', tags: ['reasoning', 'pro'], context: 1050000, priceIn: 30, priceOut: 180 },
+
+  // ── GPT-5 Codex (coding) ──
+  { id: 'gpt-5.3-codex', name: 'GPT-5.3 Codex', provider: 'openai', tags: ['coding', 'reasoning'], context: 400000, priceIn: 1.75, priceOut: 14 },
+  { id: 'gpt-5.3-codex-spark', name: 'GPT-5.3 Codex Spark', provider: 'openai', tags: ['coding', 'fast', 'reasoning'], context: 128000, priceIn: 1.75, priceOut: 14 },
+
+  // ── GPT-5 Series ──
+  { id: 'gpt-5.2', name: 'GPT-5.2', provider: 'openai', tags: ['reasoning'], context: 400000, priceIn: 1.75, priceOut: 14 },
+  { id: 'gpt-5.1', name: 'GPT-5.1', provider: 'openai', tags: ['reasoning'], context: 400000, priceIn: 1.25, priceOut: 10 },
+  { id: 'gpt-5', name: 'GPT-5', provider: 'openai', tags: ['frontier', 'reasoning'], context: 400000, priceIn: 1.25, priceOut: 10 },
+  { id: 'gpt-5-mini', name: 'GPT-5 Mini', provider: 'openai', tags: ['fast', 'reasoning'], context: 400000, priceIn: 0.25, priceOut: 2 },
+  { id: 'gpt-5-nano', name: 'GPT-5 Nano', provider: 'openai', tags: ['cheap', 'fast', 'reasoning'], context: 400000, priceIn: 0.05, priceOut: 0.4 },
+  { id: 'gpt-5-pro', name: 'GPT-5 Pro', provider: 'openai', tags: ['reasoning', 'pro'], context: 400000, priceIn: 15, priceOut: 120 },
+
+  // ── O-Series (Reasoning, legacy) ──
+  { id: 'o3', name: 'o3', provider: 'openai', tags: ['reasoning', 'legacy'], context: 200000, priceIn: 2, priceOut: 8 },
+  { id: 'o3-pro', name: 'o3 Pro', provider: 'openai', tags: ['reasoning', 'pro', 'legacy'], context: 200000, priceIn: 20, priceOut: 80 },
+  { id: 'o3-mini', name: 'o3 Mini', provider: 'openai', tags: ['reasoning', 'fast', 'legacy'], context: 200000, priceIn: 1.1, priceOut: 4.4 },
+  { id: 'o4-mini', name: 'o4 Mini', provider: 'openai', tags: ['reasoning', 'fast', 'legacy'], context: 200000, priceIn: 1.1, priceOut: 4.4 },
+
+  // ── GPT-4.1 / 4o (legacy) ──
+  { id: 'gpt-4.1', name: 'GPT-4.1', provider: 'openai', tags: ['legacy'], context: 1047576, priceIn: 2, priceOut: 8 },
+  { id: 'gpt-4.1-mini', name: 'GPT-4.1 Mini', provider: 'openai', tags: ['fast', 'legacy'], context: 1047576, priceIn: 0.4, priceOut: 1.6 },
+  { id: 'gpt-4.1-nano', name: 'GPT-4.1 Nano', provider: 'openai', tags: ['cheap', 'fast', 'legacy'], context: 1047576, priceIn: 0.1, priceOut: 0.4 },
+  { id: 'gpt-4o', name: 'GPT-4o', provider: 'openai', tags: ['legacy', 'multimodal'], context: 128000, priceIn: 2.5, priceOut: 10 },
+  { id: 'gpt-4o-mini', name: 'GPT-4o Mini', provider: 'openai', tags: ['fast', 'multimodal', 'legacy'], context: 128000, priceIn: 0.15, priceOut: 0.6 },
 ],
 };
 

package/src/providers/openrouter.js

@@ -12,32 +12,62 @@ const provider = {
     envKeyHint: 'OPENROUTER',
   }),
   models: [
-  // Anthropic Models
-  { id: 'anthropic/claude-3-opus', name: 'Claude 3 Opus (via OpenRouter)', provider: 'openrouter', context: 200000, priceIn: 15, priceOut: 75 },
-  
-  // OpenAI Models
-  { id: 'openai/gpt-4o-mini', name: 'GPT-4o Mini (via OpenRouter)', provider: 'openrouter', context: 128000, priceIn: 0.15, priceOut: 0.60 },
-  { id: 'openai/gpt-4o', name: 'GPT-4o (via OpenRouter)', provider: 'openrouter', context: 128000, priceIn: 2.50, priceOut: 10.00 },
-  { id: 'openai/o3-mini-high', name: 'O3 Mini High (via OpenRouter)', provider: 'openrouter', context: 200000, priceIn: 1.10, priceOut: 4.40 },
-  { id: 'openai/o1', name: 'O1 (via OpenRouter)', provider: 'openrouter', context: 200000, priceIn: 15, priceOut: 60 },
-  
-  // Google Models
-  { id: 'google/gemini-2.5-pro', name: 'Gemini 2.5 Pro (via OpenRouter)', provider: 'openrouter', context: 1048576, priceIn: 1.25, priceOut: 10 },
-  { id: 'google/gemini-2.5-flash', name: 'Gemini 2.5 Flash (via OpenRouter)', provider: 'openrouter', context: 1048576, priceIn: 0.15, priceOut: 1.25 },
-  
-  // DeepSeek
-  { id: 'deepseek/deepseek-v4-flash', name: 'DeepSeek V4 Flash (via OpenRouter)', provider: 'openrouter', context: 1048576, priceIn: 0.07, priceOut: 0.18 },
-  
-  // NVIDIA free-tier variants (the :free slugs are genuinely free on
-  // OpenRouter; the un-suffixed Nemotron IDs are paid and must not be tagged free)
-  { id: 'nvidia/nemotron-3-super-120b-a12b:free', name: 'Nemotron 3 Super (free, via OpenRouter)', provider: 'openrouter', tags: ['free'], context: 262144, priceIn: 0, priceOut: 0 },
-  { id: 'nvidia/nemotron-3-ultra-550b-a55b:free', name: 'Nemotron 3 Ultra (free, via OpenRouter)', provider: 'openrouter', tags: ['free'], context: 1048576, priceIn: 0, priceOut: 0 },
-  
-  // Other Top-Tier Alternatives
-  { id: 'zhipuai/glm-4-plus', name: 'GLM-4 Plus (via OpenRouter)', provider: 'openrouter', context: 128000, priceIn: 1.5, priceOut: 1.5 },
-  { id: 'minimax/minimax-01', name: 'MiniMax-01 (via OpenRouter)', provider: 'openrouter', context: 1000000, priceIn: 0.14, priceOut: 1.4 },
-  { id: 'mistralai/mistral-large-2411', name: 'Mistral Large 2411 (via OpenRouter)', provider: 'openrouter', context: 128000, priceIn: 2, priceOut: 6 }
-]
+  // ── Anthropic ──
+  { id: 'anthropic/claude-opus-5.5', name: 'Claude Opus 5.5 (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1000000, priceIn: 4, priceOut: 20 },
+  { id: 'anthropic/claude-opus-4.8', name: 'Claude Opus 4.8 (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1000000, priceIn: 5, priceOut: 25 },
+  { id: 'anthropic/claude-sonnet-5.5', name: 'Claude Sonnet 5.5 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1000000, priceIn: 2, priceOut: 10 },
+  { id: 'anthropic/claude-sonnet-4.6', name: 'Claude Sonnet 4.6 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1000000, priceIn: 3, priceOut: 15 },
+  { id: 'anthropic/claude-haiku-5.5', name: 'Claude Haiku 5.5 (via OpenRouter)', provider: 'openrouter', tags: ['fast', 'reasoning'], context: 1000000, priceIn: 0.1, priceOut: 0.5 },
+
+  // ── OpenAI ──
+  { id: 'openai/gpt-6-astra', name: 'GPT-6 Astra (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1050000, priceIn: 10, priceOut: 50 },
+  { id: 'openai/gpt-6-sol', name: 'GPT-6 Sol (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1050000, priceIn: 2, priceOut: 10 },
+  { id: 'openai/gpt-6-luna', name: 'GPT-6 Luna (via OpenRouter)', provider: 'openrouter', tags: ['fast', 'cheap', 'reasoning'], context: 1050000, priceIn: 0.1, priceOut: 0.5 },
+  { id: 'openai/gpt-5.6-sol', name: 'GPT-5.6 Sol (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1050000, priceIn: 4, priceOut: 20 },
+  { id: 'openai/gpt-5.6-terra', name: 'GPT-5.6 Terra (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1050000, priceIn: 2, priceOut: 12 },
+  { id: 'openai/gpt-5.6-luna', name: 'GPT-5.6 Luna (via OpenRouter)', provider: 'openrouter', tags: ['fast', 'cheap', 'reasoning'], context: 1050000, priceIn: 0.2, priceOut: 1.2 },
+  { id: 'openai/gpt-5.5', name: 'GPT-5.5 (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1050000, priceIn: 5, priceOut: 30 },
+  { id: 'openai/gpt-5.4', name: 'GPT-5.4 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1050000, priceIn: 2.5, priceOut: 15 },
+  { id: 'openai/gpt-5.3-codex', name: 'GPT-5.3 Codex (via OpenRouter)', provider: 'openrouter', tags: ['coding', 'reasoning'], context: 400000, priceIn: 1.75, priceOut: 14 },
+  { id: 'openai/o3', name: 'o3 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning', 'legacy'], context: 200000, priceIn: 2, priceOut: 8 },
+
+  // ── Google ──
+  { id: 'google/gemini-4-argon', name: 'Gemini 4 Argon (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1048576, priceIn: 2, priceOut: 10 },
+  { id: 'google/gemini-3.1-pro-preview', name: 'Gemini 3.1 Pro Preview (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1048576, priceIn: 2, priceOut: 12 },
+  { id: 'google/gemini-3.8-flash', name: 'Gemini 3.8 Flash (via OpenRouter)', provider: 'openrouter', tags: ['fast', 'reasoning'], context: 1048576, priceIn: 0.82, priceOut: 4.12 },
+  { id: 'google/gemini-3.5-flash', name: 'Gemini 3.5 Flash (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1048576, priceIn: 1.5, priceOut: 9 },
+
+  // ── DeepSeek ──
+  { id: 'deepseek/deepseek-v4.1-flash', name: 'DeepSeek V4.1 Flash (via OpenRouter)', provider: 'openrouter', tags: ['fast', 'reasoning'], context: 1048576, priceIn: 0.04, priceOut: 0.64 },
+  { id: 'deepseek/deepseek-v4-pro', name: 'DeepSeek V4 Pro (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1048576, priceIn: 0.3, priceOut: 0.59 },
+
+  // ── Moonshot / Z.ai / Qwen / MiniMax ──
+  { id: 'moonshotai/kimi-k3', name: 'Kimi K3 (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 1048576, priceIn: 0.3, priceOut: 11.94 },
+  { id: 'moonshotai/kimi-k2.6', name: 'Kimi K2.6 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 262144, priceIn: 0.65, priceOut: 3.41 },
+  { id: 'z-ai/glm-5.3', name: 'GLM-5.3 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1048576, priceIn: 0.08, priceOut: 4.8 },
+  { id: 'z-ai/glm-5.3-flash', name: 'GLM-5.3 Flash (via OpenRouter)', provider: 'openrouter', tags: ['fast', 'reasoning'], context: 1048576, priceIn: 0.03, priceOut: 0.58 },
+  { id: 'qwen/qwen3.5-397b-a17b', name: 'Qwen3.5 397B A17B (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 262144, priceIn: 0.5, priceOut: 3.6 },
+  { id: 'qwen/qwen3-coder', name: 'Qwen3 Coder 480B (via OpenRouter)', provider: 'openrouter', tags: ['coding'], context: 262144, priceIn: 0.3, priceOut: 1 },
+  { id: 'qwen/qwen3-coder-flash', name: 'Qwen3 Coder Flash (via OpenRouter)', provider: 'openrouter', tags: ['coding', 'fast'], context: 1000000, priceIn: 0.2, priceOut: 0.98 },
+  { id: 'minimax/minimax-m3', name: 'MiniMax M3 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1048576, priceIn: 0.3, priceOut: 1.2 },
+
+  // ── xAI / Meta / Mistral / Cohere / Writer ──
+  { id: 'x-ai/grok-4.7', name: 'Grok 4.7 (via OpenRouter)', provider: 'openrouter', tags: ['frontier', 'reasoning'], context: 500000, priceIn: 2, priceOut: 6 },
+  { id: 'x-ai/grok-4.3', name: 'Grok 4.3 (via OpenRouter)', provider: 'openrouter', tags: ['reasoning'], context: 1000000, priceIn: 1.25, priceOut: 2.5 },
+  { id: 'meta-llama/llama-4-maverick', name: 'Llama 4 Maverick (via OpenRouter)', provider: 'openrouter', tags: [], context: 1048576, priceIn: 0.19, priceOut: 0.65 },
+  { id: 'meta-llama/llama-4-scout', name: 'Llama 4 Scout (via OpenRouter)', provider: 'openrouter', tags: ['fast'], context: 1310720, priceIn: 0.1, priceOut: 0.3 },
+  { id: 'mistralai/mistral-large-4-0', name: 'Mistral Large 4 (via OpenRouter)', provider: 'openrouter', tags: ['frontier'], context: 1048576, priceIn: 0.68, priceOut: 2.09 },
+  { id: 'cohere/command-a', name: 'Command A (via OpenRouter)', provider: 'openrouter', tags: [], context: 256000, priceIn: 2.5, priceOut: 10 },
+  { id: 'writer/palmyra-x5', name: 'Palmyra X5 (via OpenRouter)', provider: 'openrouter', tags: [], context: 1040000, priceIn: 0.6, priceOut: 6 },
+
+  // ── Free tiers (genuinely free slugs) ──
+  { id: 'nvidia/nemotron-3-super-120b-a12b:free', name: 'Nemotron 3 Super (free, via OpenRouter)', provider: 'openrouter', tags: ['free', 'reasoning'], context: 262144, priceIn: 0, priceOut: 0 },
+  { id: 'nvidia/nemotron-3-ultra-550b-a55b:free', name: 'Nemotron 3 Ultra (free, via OpenRouter)', provider: 'openrouter', tags: ['free', 'reasoning'], context: 1000000, priceIn: 0, priceOut: 0 },
+  { id: 'nvidia/nemotron-3.5-lightning:free', name: 'Nemotron 3.5 Lightning (free, via OpenRouter)', provider: 'openrouter', tags: ['free', 'fast', 'reasoning'], context: 1000000, priceIn: 0, priceOut: 0 },
+  { id: 'nvidia/nemotron-3.5-super-vl:free', name: 'Nemotron 3.5 Super VL (free, via OpenRouter)', provider: 'openrouter', tags: ['free', 'vision', 'reasoning'], context: 1000000, priceIn: 0, priceOut: 0 },
+  { id: 'google/gemma-4-31b-it:free', name: 'Gemma 4 31B (free, via OpenRouter)', provider: 'openrouter', tags: ['free', 'reasoning'], context: 262144, priceIn: 0, priceOut: 0 },
+  { id: 'google/gemma-4-26b-a4b-it:free', name: 'Gemma 4 26B A4B (free, via OpenRouter)', provider: 'openrouter', tags: ['free', 'fast', 'reasoning'], context: 262144, priceIn: 0, priceOut: 0 },
+],
 };
 
 module.exports = provider;