mcp-cost-tracker 1.0.0

package/src/pricing/models.ts

@@ -0,0 +1,246 @@
+/**
+ * LLM Model Pricing Database
+ * Prices are in USD per 1 million tokens
+ * Last updated: February 2026
+ */
+
+export interface ModelPricing {
+  provider: string;
+  model: string;
+  inputPricePerMTok: number;
+  outputPricePerMTok: number;
+  cachedInputPricePerMTok?: number;
+  contextWindow?: number;
+  category: 'chat' | 'reasoning' | 'embedding' | 'image' | 'audio' | 'code';
+  notes?: string;
+}
+
+/**
+ * Comprehensive pricing database for all major LLM providers.
+ * Prices are per 1 million tokens (MTok) in USD.
+ */
+export const MODEL_PRICING: ModelPricing[] = [
+  // ============================================================
+  // OpenAI Models
+  // ============================================================
+  // GPT-4.1 Series
+  { provider: 'openai', model: 'gpt-4.1', inputPricePerMTok: 2.00, outputPricePerMTok: 8.00, cachedInputPricePerMTok: 0.50, contextWindow: 1047576, category: 'chat' },
+  { provider: 'openai', model: 'gpt-4.1-mini', inputPricePerMTok: 0.40, outputPricePerMTok: 1.60, cachedInputPricePerMTok: 0.10, contextWindow: 1047576, category: 'chat' },
+  { provider: 'openai', model: 'gpt-4.1-nano', inputPricePerMTok: 0.10, outputPricePerMTok: 0.40, cachedInputPricePerMTok: 0.025, contextWindow: 1047576, category: 'chat' },
+
+  // GPT-4o Series
+  { provider: 'openai', model: 'gpt-4o', inputPricePerMTok: 2.50, outputPricePerMTok: 10.00, cachedInputPricePerMTok: 1.25, contextWindow: 128000, category: 'chat' },
+  { provider: 'openai', model: 'gpt-4o-mini', inputPricePerMTok: 0.15, outputPricePerMTok: 0.60, cachedInputPricePerMTok: 0.075, contextWindow: 128000, category: 'chat' },
+
+  // GPT-5 Series
+  { provider: 'openai', model: 'gpt-5', inputPricePerMTok: 10.00, outputPricePerMTok: 30.00, cachedInputPricePerMTok: 2.50, contextWindow: 128000, category: 'chat' },
+  { provider: 'openai', model: 'gpt-5-mini', inputPricePerMTok: 1.50, outputPricePerMTok: 6.00, cachedInputPricePerMTok: 0.375, contextWindow: 128000, category: 'chat' },
+  { provider: 'openai', model: 'gpt-5-nano', inputPricePerMTok: 0.05, outputPricePerMTok: 0.40, contextWindow: 128000, category: 'chat' },
+
+  // O-Series (Reasoning)
+  { provider: 'openai', model: 'o1', inputPricePerMTok: 15.00, outputPricePerMTok: 60.00, cachedInputPricePerMTok: 7.50, contextWindow: 200000, category: 'reasoning' },
+  { provider: 'openai', model: 'o1-mini', inputPricePerMTok: 1.10, outputPricePerMTok: 4.40, cachedInputPricePerMTok: 0.55, contextWindow: 128000, category: 'reasoning' },
+  { provider: 'openai', model: 'o1-pro', inputPricePerMTok: 150.00, outputPricePerMTok: 600.00, contextWindow: 200000, category: 'reasoning' },
+  { provider: 'openai', model: 'o3', inputPricePerMTok: 2.00, outputPricePerMTok: 8.00, cachedInputPricePerMTok: 0.50, contextWindow: 200000, category: 'reasoning' },
+  { provider: 'openai', model: 'o3-mini', inputPricePerMTok: 1.10, outputPricePerMTok: 4.40, cachedInputPricePerMTok: 0.55, contextWindow: 200000, category: 'reasoning' },
+  { provider: 'openai', model: 'o4-mini', inputPricePerMTok: 1.10, outputPricePerMTok: 4.40, cachedInputPricePerMTok: 0.55, contextWindow: 200000, category: 'reasoning' },
+
+  // Legacy GPT-4
+  { provider: 'openai', model: 'gpt-4-turbo', inputPricePerMTok: 10.00, outputPricePerMTok: 30.00, contextWindow: 128000, category: 'chat' },
+  { provider: 'openai', model: 'gpt-4', inputPricePerMTok: 30.00, outputPricePerMTok: 60.00, contextWindow: 8192, category: 'chat' },
+  { provider: 'openai', model: 'gpt-3.5-turbo', inputPricePerMTok: 0.50, outputPricePerMTok: 1.50, contextWindow: 16385, category: 'chat' },
+
+  // Embeddings
+  { provider: 'openai', model: 'text-embedding-3-large', inputPricePerMTok: 0.13, outputPricePerMTok: 0, contextWindow: 8191, category: 'embedding' },
+  { provider: 'openai', model: 'text-embedding-3-small', inputPricePerMTok: 0.02, outputPricePerMTok: 0, contextWindow: 8191, category: 'embedding' },
+  { provider: 'openai', model: 'text-embedding-ada-002', inputPricePerMTok: 0.10, outputPricePerMTok: 0, contextWindow: 8191, category: 'embedding' },
+
+  // ============================================================
+  // Anthropic Models
+  // ============================================================
+  { provider: 'anthropic', model: 'claude-opus-4', inputPricePerMTok: 15.00, outputPricePerMTok: 75.00, cachedInputPricePerMTok: 1.875, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-sonnet-4', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, cachedInputPricePerMTok: 0.375, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-sonnet-4.5', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, cachedInputPricePerMTok: 0.375, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-sonnet-4-5', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, cachedInputPricePerMTok: 0.375, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-5-sonnet-20241022', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, cachedInputPricePerMTok: 0.375, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-5-sonnet-latest', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, cachedInputPricePerMTok: 0.375, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-5-haiku-20241022', inputPricePerMTok: 0.80, outputPricePerMTok: 4.00, cachedInputPricePerMTok: 0.08, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-haiku-20240307', inputPricePerMTok: 0.25, outputPricePerMTok: 1.25, cachedInputPricePerMTok: 0.03, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-opus-20240229', inputPricePerMTok: 15.00, outputPricePerMTok: 75.00, cachedInputPricePerMTok: 1.875, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-sonnet-20240229', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, contextWindow: 200000, category: 'chat' },
+
+  // Anthropic aliases
+  { provider: 'anthropic', model: 'claude-3.5-sonnet', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, cachedInputPricePerMTok: 0.375, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3.5-haiku', inputPricePerMTok: 0.80, outputPricePerMTok: 4.00, cachedInputPricePerMTok: 0.08, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-opus', inputPricePerMTok: 15.00, outputPricePerMTok: 75.00, cachedInputPricePerMTok: 1.875, contextWindow: 200000, category: 'chat' },
+  { provider: 'anthropic', model: 'claude-3-haiku', inputPricePerMTok: 0.25, outputPricePerMTok: 1.25, cachedInputPricePerMTok: 0.03, contextWindow: 200000, category: 'chat' },
+
+  // ============================================================
+  // Google Models
+  // ============================================================
+  { provider: 'google', model: 'gemini-2.5-pro', inputPricePerMTok: 1.25, outputPricePerMTok: 10.00, contextWindow: 1048576, category: 'chat', notes: 'Price for ≤200k tokens' },
+  { provider: 'google', model: 'gemini-2.5-flash', inputPricePerMTok: 0.15, outputPricePerMTok: 0.60, contextWindow: 1048576, category: 'chat' },
+  { provider: 'google', model: 'gemini-2.0-flash', inputPricePerMTok: 0.10, outputPricePerMTok: 0.40, contextWindow: 1048576, category: 'chat' },
+  { provider: 'google', model: 'gemini-2.0-flash-lite', inputPricePerMTok: 0.075, outputPricePerMTok: 0.30, contextWindow: 1048576, category: 'chat' },
+  { provider: 'google', model: 'gemini-1.5-pro', inputPricePerMTok: 1.25, outputPricePerMTok: 5.00, contextWindow: 2097152, category: 'chat' },
+  { provider: 'google', model: 'gemini-1.5-flash', inputPricePerMTok: 0.075, outputPricePerMTok: 0.30, contextWindow: 1048576, category: 'chat' },
+  { provider: 'google', model: 'gemini-1.0-pro', inputPricePerMTok: 0.50, outputPricePerMTok: 1.50, contextWindow: 32760, category: 'chat' },
+
+  // ============================================================
+  // DeepSeek Models
+  // ============================================================
+  { provider: 'deepseek', model: 'deepseek-chat', inputPricePerMTok: 0.27, outputPricePerMTok: 1.10, cachedInputPricePerMTok: 0.07, contextWindow: 65536, category: 'chat' },
+  { provider: 'deepseek', model: 'deepseek-v3', inputPricePerMTok: 0.27, outputPricePerMTok: 1.10, cachedInputPricePerMTok: 0.07, contextWindow: 65536, category: 'chat' },
+  { provider: 'deepseek', model: 'deepseek-reasoner', inputPricePerMTok: 0.55, outputPricePerMTok: 2.19, cachedInputPricePerMTok: 0.14, contextWindow: 65536, category: 'reasoning' },
+  { provider: 'deepseek', model: 'deepseek-r1', inputPricePerMTok: 0.55, outputPricePerMTok: 2.19, cachedInputPricePerMTok: 0.14, contextWindow: 65536, category: 'reasoning' },
+
+  // ============================================================
+  // Mistral Models
+  // ============================================================
+  { provider: 'mistral', model: 'mistral-large-latest', inputPricePerMTok: 2.00, outputPricePerMTok: 6.00, contextWindow: 131072, category: 'chat' },
+  { provider: 'mistral', model: 'mistral-medium-latest', inputPricePerMTok: 0.40, outputPricePerMTok: 2.00, contextWindow: 131072, category: 'chat' },
+  { provider: 'mistral', model: 'mistral-small-latest', inputPricePerMTok: 0.10, outputPricePerMTok: 0.30, contextWindow: 131072, category: 'chat' },
+  { provider: 'mistral', model: 'codestral-latest', inputPricePerMTok: 0.30, outputPricePerMTok: 0.90, contextWindow: 262144, category: 'code' },
+  { provider: 'mistral', model: 'pixtral-large-latest', inputPricePerMTok: 2.00, outputPricePerMTok: 6.00, contextWindow: 131072, category: 'chat' },
+  { provider: 'mistral', model: 'mistral-nemo', inputPricePerMTok: 0.15, outputPricePerMTok: 0.15, contextWindow: 131072, category: 'chat' },
+
+  // ============================================================
+  // Meta Llama Models (via common providers)
+  // ============================================================
+  { provider: 'meta', model: 'llama-3.3-70b', inputPricePerMTok: 0.18, outputPricePerMTok: 0.18, contextWindow: 131072, category: 'chat' },
+  { provider: 'meta', model: 'llama-3.1-405b', inputPricePerMTok: 0.80, outputPricePerMTok: 0.80, contextWindow: 131072, category: 'chat' },
+  { provider: 'meta', model: 'llama-3.1-70b', inputPricePerMTok: 0.18, outputPricePerMTok: 0.18, contextWindow: 131072, category: 'chat' },
+  { provider: 'meta', model: 'llama-3.1-8b', inputPricePerMTok: 0.02, outputPricePerMTok: 0.05, contextWindow: 131072, category: 'chat' },
+  { provider: 'meta', model: 'llama-4-scout', inputPricePerMTok: 0.15, outputPricePerMTok: 0.60, contextWindow: 512000, category: 'chat' },
+  { provider: 'meta', model: 'llama-4-maverick', inputPricePerMTok: 0.30, outputPricePerMTok: 0.90, contextWindow: 1048576, category: 'chat' },
+
+  // ============================================================
+  // Cohere Models
+  // ============================================================
+  { provider: 'cohere', model: 'command-r-plus', inputPricePerMTok: 2.50, outputPricePerMTok: 10.00, contextWindow: 128000, category: 'chat' },
+  { provider: 'cohere', model: 'command-r', inputPricePerMTok: 0.15, outputPricePerMTok: 0.60, contextWindow: 128000, category: 'chat' },
+  { provider: 'cohere', model: 'command-light', inputPricePerMTok: 0.30, outputPricePerMTok: 0.60, contextWindow: 4096, category: 'chat' },
+
+  // ============================================================
+  // xAI Models
+  // ============================================================
+  { provider: 'xai', model: 'grok-3', inputPricePerMTok: 3.00, outputPricePerMTok: 15.00, contextWindow: 131072, category: 'chat' },
+  { provider: 'xai', model: 'grok-3-mini', inputPricePerMTok: 0.30, outputPricePerMTok: 0.50, contextWindow: 131072, category: 'reasoning' },
+  { provider: 'xai', model: 'grok-2', inputPricePerMTok: 2.00, outputPricePerMTok: 10.00, contextWindow: 131072, category: 'chat' },
+];
+
+/**
+ * Build a lookup map for fast model price resolution.
+ * Supports fuzzy matching by normalizing model names.
+ */
+const pricingMap = new Map<string, ModelPricing>();
+
+function normalizeModelName(name: string): string {
+  return name
+    .toLowerCase()
+    .replace(/[_\s]/g, '-')
+    .replace(/^(models\/|ft:|accounts\/[^/]+\/models\/)/, '')
+    .trim();
+}
+
+for (const pricing of MODEL_PRICING) {
+  pricingMap.set(normalizeModelName(pricing.model), pricing);
+  // Also index with provider prefix
+  pricingMap.set(`${pricing.provider}/${normalizeModelName(pricing.model)}`, pricing);
+}
+
+/**
+ * Look up pricing for a given model name.
+ * Supports exact match, provider-prefixed match, and fuzzy partial match.
+ */
+export function lookupPricing(model: string, provider?: string): ModelPricing | null {
+  const normalized = normalizeModelName(model);
+
+  // 1. Try exact match with provider prefix
+  if (provider) {
+    const key = `${provider.toLowerCase()}/${normalized}`;
+    const exact = pricingMap.get(key);
+    if (exact) return exact;
+  }
+
+  // 2. Try exact match without provider
+  const exact = pricingMap.get(normalized);
+  if (exact) return exact;
+
+  // 3. Fuzzy match: find the best partial match
+  let bestMatch: ModelPricing | null = null;
+  let bestScore = 0;
+
+  for (const pricing of MODEL_PRICING) {
+    const pricingNorm = normalizeModelName(pricing.model);
+
+    // Check if the query contains the model name or vice versa
+    if (normalized.includes(pricingNorm) || pricingNorm.includes(normalized)) {
+      const score = pricingNorm.length;
+      // If provider matches, boost score
+      const providerBoost = provider && pricing.provider === provider.toLowerCase() ? 1000 : 0;
+      if (score + providerBoost > bestScore) {
+        bestScore = score + providerBoost;
+        bestMatch = pricing;
+      }
+    }
+  }
+
+  return bestMatch;
+}
+
+/**
+ * Calculate cost for a given number of tokens.
+ */
+export function calculateCost(
+  model: string,
+  inputTokens: number,
+  outputTokens: number,
+  provider?: string,
+  cachedInputTokens?: number
+): { cost: number; inputCost: number; outputCost: number; cachedInputCost: number; pricing: ModelPricing | null } {
+  const pricing = lookupPricing(model, provider);
+
+  if (!pricing) {
+    return { cost: 0, inputCost: 0, outputCost: 0, cachedInputCost: 0, pricing: null };
+  }
+
+  const effectiveInputTokens = cachedInputTokens
+    ? inputTokens - cachedInputTokens
+    : inputTokens;
+
+  const inputCost = (effectiveInputTokens / 1_000_000) * pricing.inputPricePerMTok;
+  const outputCost = (outputTokens / 1_000_000) * pricing.outputPricePerMTok;
+  const cachedInputCost = cachedInputTokens && pricing.cachedInputPricePerMTok
+    ? (cachedInputTokens / 1_000_000) * pricing.cachedInputPricePerMTok
+    : 0;
+
+  const cost = inputCost + outputCost + cachedInputCost;
+
+  return { cost, inputCost, outputCost, cachedInputCost, pricing };
+}
+
+/**
+ * Get all supported providers.
+ */
+export function getSupportedProviders(): string[] {
+  const providers = new Set<string>();
+  for (const pricing of MODEL_PRICING) {
+    providers.add(pricing.provider);
+  }
+  return Array.from(providers).sort();
+}
+
+/**
+ * Get all models for a given provider.
+ */
+export function getModelsForProvider(provider: string): ModelPricing[] {
+  return MODEL_PRICING.filter(p => p.provider === provider.toLowerCase());
+}
+
+/**
+ * Get all available models.
+ */
+export function getAllModels(): ModelPricing[] {
+  return [...MODEL_PRICING];
+}