/** * Static defaults for well-known models: capabilities, limits and list prices. * * Discovery APIs return ids (and sometimes limits) but almost never prices or * feature flags, so `listModels` merges its results with this table. Unknown * models simply get no defaults — nothing here is required for a model to work. * * Everything below was read from the official pages listed per provider on the * dates noted. Prices are USD per million tokens, standard (non-batch, * non-priority) tier, short-context tier where a provider has several. A field * is omitted when it could not be confirmed from the page — do not guess. */ import type { DiscoveredModel, ModelCapabilities, ProviderKind } from '../types.ts'; export const CATALOG_UPDATED = '2026-11-10'; type Entry = Omit & { capabilities?: Partial }; /* ------------------------------------------------------------------ */ /* OpenAI */ /* Prices: https://developers.openai.com/api/docs/pricing (fetched 2026-11-11) /* Limits: https://developers.openai.com/api/docs/models (fetched 2026-11-11; /* only the four current flagship models list limits there — the /* limits on older models below are the long-published values and were /* not re-fetched) /* Reasoning models reject a custom `temperature`, hence temperature:false. /* ------------------------------------------------------------------ */ const OA_REASONING: Partial = { tools: true, vision: true, pdf: true, reasoning: true, temperature: false }; const OA_CLASSIC: Partial = { tools: true, vision: true, pdf: true, reasoning: false, temperature: true }; const OA_FLAGSHIP_LIMITS = { contextWindow: 1_050_000, maxOutput: 128_000 }; const OPENAI: Record = { 'gpt-6-astra': { label: 'GPT-6 Astra', ...OA_FLAGSHIP_LIMITS, priceIn: 10, priceOut: 50, capabilities: OA_REASONING }, 'gpt-6.1-sol': { label: 'GPT-6.1 Sol', ...OA_FLAGSHIP_LIMITS, priceIn: 2, priceOut: 10, capabilities: OA_REASONING }, 'gpt-6.1-terra': { label: 'GPT-6.1 Terra', ...OA_FLAGSHIP_LIMITS, priceIn: 1, priceOut: 5, capabilities: OA_REASONING }, 'gpt-6-luna': { label: 'GPT-6 Luna', ...OA_FLAGSHIP_LIMITS, priceIn: 0.1, priceOut: 0.5, capabilities: OA_REASONING }, 'gpt-6-sol': { label: 'GPT-6 Sol', priceIn: 2, priceOut: 10, capabilities: OA_REASONING }, 'gpt-5.6-sol': { label: 'GPT-5.6 Sol', priceIn: 4, priceOut: 20, capabilities: OA_REASONING }, 'gpt-5.6-terra': { label: 'GPT-5.6 Terra', priceIn: 2, priceOut: 12, capabilities: OA_REASONING }, 'gpt-5.6-luna': { label: 'GPT-5.6 Luna', priceIn: 0.2, priceOut: 1.2, capabilities: OA_REASONING }, 'gpt-5.5': { label: 'GPT-5.5', priceIn: 5, priceOut: 30, capabilities: OA_REASONING }, 'gpt-5.5-pro': { label: 'GPT-5.5 Pro', priceIn: 30, priceOut: 180, capabilities: OA_REASONING }, 'gpt-5.4': { label: 'GPT-5.4', priceIn: 2.5, priceOut: 15, capabilities: OA_REASONING }, 'gpt-5.4-pro': { label: 'GPT-5.4 Pro', priceIn: 30, priceOut: 180, capabilities: OA_REASONING }, 'gpt-5.4-mini': { label: 'GPT-5.4 mini', priceIn: 0.75, priceOut: 4.5, capabilities: OA_REASONING }, 'gpt-5.4-nano': { label: 'GPT-5.4 nano', priceIn: 0.2, priceOut: 1.25, capabilities: OA_REASONING }, 'gpt-5.3-codex': { label: 'GPT-5.3 Codex', priceIn: 1.75, priceOut: 14, capabilities: OA_REASONING }, 'gpt-5.2': { label: 'GPT-5.2', priceIn: 1.75, priceOut: 14, capabilities: OA_REASONING }, 'gpt-5.2-pro': { label: 'GPT-5.2 Pro', priceIn: 21, priceOut: 168, capabilities: OA_REASONING }, 'gpt-5.1': { label: 'GPT-5.1', priceIn: 1.25, priceOut: 10, capabilities: OA_REASONING }, 'gpt-5': { label: 'GPT-5', contextWindow: 400_000, maxOutput: 128_000, priceIn: 1.25, priceOut: 10, capabilities: OA_REASONING }, 'gpt-5-mini': { label: 'GPT-5 mini', contextWindow: 400_000, maxOutput: 128_000, priceIn: 0.25, priceOut: 2, capabilities: OA_REASONING }, 'gpt-5-nano': { label: 'GPT-5 nano', contextWindow: 400_000, maxOutput: 128_000, priceIn: 0.05, priceOut: 0.4, capabilities: OA_REASONING }, 'gpt-5-pro': { label: 'GPT-5 Pro', priceIn: 15, priceOut: 120, capabilities: OA_REASONING }, 'chat-latest': { label: 'ChatGPT (latest)', priceIn: 5, priceOut: 30, capabilities: OA_CLASSIC }, 'gpt-4.1': { label: 'GPT-4.1', contextWindow: 1_047_576, maxOutput: 32_768, priceIn: 2, priceOut: 8, capabilities: OA_CLASSIC }, 'gpt-4.1-mini': { label: 'GPT-4.1 mini', contextWindow: 1_047_576, maxOutput: 32_768, priceIn: 0.4, priceOut: 1.6, capabilities: OA_CLASSIC }, 'gpt-4o': { label: 'GPT-4o', contextWindow: 128_000, maxOutput: 16_384, priceIn: 2.5, priceOut: 10, capabilities: OA_CLASSIC }, 'gpt-4o-mini': { label: 'GPT-4o mini', contextWindow: 128_000, maxOutput: 16_384, priceIn: 0.15, priceOut: 0.6, capabilities: OA_CLASSIC }, o3: { label: 'o3', contextWindow: 200_000, maxOutput: 100_000, priceIn: 2, priceOut: 8, capabilities: OA_REASONING }, 'o3-pro': { label: 'o3-pro', contextWindow: 200_000, maxOutput: 100_000, priceIn: 20, priceOut: 80, capabilities: OA_REASONING }, }; /* ------------------------------------------------------------------ */ /* Anthropic */ /* Prices: https://platform.claude.com/docs/en/about-claude/pricing (fetched 2026-11-11) /* Ids and limits: https://platform.claude.com/docs/en/about-claude/models/overview /* (via Anthropic's Claude API reference, cached 2026-11-03) /* temperature:false for every Claude model: the adapter speaks API version /* 2026-10-21, which rejects temperature/top_p/top_k on all models /* (https://platform.claude.com/docs/en/api/migration-2026-10-21). /* Claude Haiku 5.5 prices are for prompts up to 100K tokens ($0.50/$2.50 above). /* ------------------------------------------------------------------ */ const CLAUDE: Partial = { tools: true, vision: true, pdf: true, reasoning: true, temperature: false }; const CLAUDE_1M = { contextWindow: 1_000_000, maxOutput: 128_000 }; const ANTHROPIC: Record = { 'claude-fable-5-2': { label: 'Claude Fable 5.2', ...CLAUDE_1M, priceIn: 8, priceOut: 40, capabilities: CLAUDE }, 'claude-mythos-5-2': { label: 'Claude Mythos 5.2', ...CLAUDE_1M, priceIn: 8, priceOut: 40, capabilities: CLAUDE }, 'claude-fable-5-1': { label: 'Claude Fable 5.1', ...CLAUDE_1M, priceIn: 10, priceOut: 50, capabilities: CLAUDE }, 'claude-mythos-5-1': { label: 'Claude Mythos 5.1', ...CLAUDE_1M, priceIn: 10, priceOut: 50, capabilities: CLAUDE }, 'claude-fable-5': { label: 'Claude Fable 5', ...CLAUDE_1M, priceIn: 10, priceOut: 50, capabilities: CLAUDE }, 'claude-mythos-5': { label: 'Claude Mythos 5', ...CLAUDE_1M, priceIn: 10, priceOut: 50, capabilities: CLAUDE }, 'claude-opus-5-5': { label: 'Claude Opus 5.5', ...CLAUDE_1M, priceIn: 4, priceOut: 20, capabilities: CLAUDE }, 'claude-opus-5': { label: 'Claude Opus 5', ...CLAUDE_1M, priceIn: 5, priceOut: 25, capabilities: CLAUDE }, 'claude-opus-4-8': { label: 'Claude Opus 4.8', ...CLAUDE_1M, priceIn: 5, priceOut: 25, capabilities: CLAUDE }, 'claude-opus-4-7': { label: 'Claude Opus 4.7', ...CLAUDE_1M, priceIn: 5, priceOut: 25, capabilities: CLAUDE }, 'claude-opus-4-6': { label: 'Claude Opus 4.6', ...CLAUDE_1M, priceIn: 5, priceOut: 25, capabilities: CLAUDE }, 'claude-opus-4-5': { label: 'Claude Opus 4.5', priceIn: 5, priceOut: 25, capabilities: CLAUDE }, 'claude-sonnet-5-5': { label: 'Claude Sonnet 5.5', ...CLAUDE_1M, priceIn: 2, priceOut: 10, capabilities: CLAUDE }, 'claude-sonnet-5': { label: 'Claude Sonnet 5', ...CLAUDE_1M, priceIn: 2, priceOut: 10, capabilities: CLAUDE }, 'claude-sonnet-4-6': { label: 'Claude Sonnet 4.6', ...CLAUDE_1M, priceIn: 3, priceOut: 15, capabilities: CLAUDE }, 'claude-sonnet-4-5': { label: 'Claude Sonnet 4.5', priceIn: 3, priceOut: 15, capabilities: CLAUDE }, 'claude-haiku-5-5': { label: 'Claude Haiku 5.5', ...CLAUDE_1M, priceIn: 0.1, priceOut: 0.5, capabilities: CLAUDE }, 'claude-haiku-4-5': { label: 'Claude Haiku 4.5', contextWindow: 200_000, maxOutput: 64_000, priceIn: 1, priceOut: 5, capabilities: CLAUDE }, }; /* ------------------------------------------------------------------ */ /* Google Gemini */ /* Prices: https://ai.google.dev/gemini-api/docs/pricing (fetched 2026-11-11, /* now served from https://aistudio.google.com/docs/pricing) /* Limits: https://ai.google.dev/gemini-api/docs/models (fetched 2026-11-11; /* only the Gemini 4 models list limits — `GET /v1beta/models` /* supplies inputTokenLimit/outputTokenLimit for the rest at runtime) /* Scheduled changes announced on the pricing page (NOT applied here): /* - gemini-4-carbon / gemini-4-argon: $4 / $20 from 2026-12-01 /* - gemini-3.8-flash / gemini-3.6-flash: $1.50 / $7.50 from 2027-01-01 /* - gemini-2.5-pro / gemini-2.5-flash are deprecated and are routed to /* gemini-3.8-flash / gemini-3.5-flash-lite (and billed as such) from 2026-12-01 /* Pro-tier prices are for prompts up to 200K tokens. /* ------------------------------------------------------------------ */ const GEMINI_CAPS: Partial = { tools: true, vision: true, pdf: true, reasoning: true, temperature: true }; const GEMINI_4_LIMITS = { contextWindow: 1_048_576, maxOutput: 65_536 }; const GEMINI: Record = { 'gemini-4-carbon': { label: 'Gemini 4 Carbon', ...GEMINI_4_LIMITS, priceIn: 2, priceOut: 10, capabilities: GEMINI_CAPS }, 'gemini-4-argon': { label: 'Gemini 4 Argon', ...GEMINI_4_LIMITS, priceIn: 2, priceOut: 10, capabilities: GEMINI_CAPS }, 'gemini-3.8-flash': { label: 'Gemini 3.8 Flash', priceIn: 0.75, priceOut: 3.75, capabilities: GEMINI_CAPS }, 'gemini-3.6-flash': { label: 'Gemini 3.6 Flash', priceIn: 0.75, priceOut: 3.75, capabilities: GEMINI_CAPS }, 'gemini-3.5-flash-lite': { label: 'Gemini 3.5 Flash-Lite', priceIn: 0.3, priceOut: 2.5, capabilities: GEMINI_CAPS }, 'gemini-3.1-pro-preview': { label: 'Gemini 3.1 Pro (preview)', priceIn: 2, priceOut: 12, capabilities: GEMINI_CAPS }, 'gemini-3.1-flash-lite': { label: 'Gemini 3.1 Flash-Lite', priceIn: 0.25, priceOut: 1.5, capabilities: GEMINI_CAPS }, 'gemini-3-flash-preview': { label: 'Gemini 3 Flash (preview)', priceIn: 0.5, priceOut: 3, capabilities: GEMINI_CAPS }, 'gemini-2.5-pro': { label: 'Gemini 2.5 Pro', priceIn: 1.25, priceOut: 10, capabilities: GEMINI_CAPS }, 'gemini-2.5-flash': { label: 'Gemini 2.5 Flash', priceIn: 0.3, priceOut: 2.5, capabilities: GEMINI_CAPS }, 'gemini-2.5-flash-lite': { label: 'Gemini 2.5 Flash-Lite', priceIn: 0.1, priceOut: 0.4, capabilities: GEMINI_CAPS }, }; /* ------------------------------------------------------------------ */ /* DeepSeek */ /* Prices and limits: https://api-docs.deepseek.com/quick_start/pricing (fetched 2026-11-15) /* Peak-hour list prices are stored. Off-peak hours (everything outside 01:00-04:00 and /* 06:00-10:00 UTC, Monday to Friday) cost half; `priceMultiplier` applies that per call. /* deepseek-v4-flash / deepseek-v4-flash-vision-exp are retired aliases served by /* deepseek-flash at the Flash price. /* ------------------------------------------------------------------ */ const DS_LIMITS = { contextWindow: 1_048_576, maxOutput: 393_216 }; const DEEPSEEK: Record = { 'deepseek-flash': { label: 'DeepSeek V4.1 Flash', ...DS_LIMITS, priceIn: 0.3, priceCachedIn: 0.006, priceOut: 1.2, capabilities: { tools: true, vision: true, pdf: false, reasoning: true, temperature: true } }, 'deepseek-v4-flash': { label: 'DeepSeek V4 Flash (alias of Flash)', ...DS_LIMITS, priceIn: 0.3, priceCachedIn: 0.006, priceOut: 1.2, capabilities: { tools: true, vision: true, pdf: false, reasoning: true, temperature: true } }, 'deepseek-v4-pro': { label: 'DeepSeek V4 Pro', ...DS_LIMITS, priceIn: 1.32, priceCachedIn: 0.044, priceOut: 3.96, capabilities: { tools: true, vision: false, pdf: false, reasoning: true, temperature: true } }, }; const TABLES: Record> = { openai: OPENAI, anthropic: ANTHROPIC, gemini: GEMINI, deepseek: DEEPSEEK, // Aggregators and local servers host arbitrary models at their own prices. openai_compatible: {}, ollama: {}, }; /** * What may follow a known id for it to still be "the same model at the same * price": a dated snapshot or a moving alias, e.g. `gpt-5-2025-08-07`, * `claude-haiku-4-5-20251001`, `claude-opus-4-5@20251101`, `gemini-2.5-flash-latest`. */ const SNAPSHOT_SUFFIX = /^[-@:](\d{4}-\d{2}-\d{2}|\d{6,8}|\d{3,4}|latest|preview|preview-\d{2}-\d{2,4}|v\d+(:\d+)?)$/; function clone(e: Entry): Partial { return { ...e, ...(e.capabilities ? { capabilities: { ...e.capabilities } } : {}) }; } function lookupIn(table: Record, id: string): Partial { const exact = table[id]; if (exact) return clone(exact); // Longest known id that is a prefix of `id` and ends on a separator. let best = ''; for (const key of Object.keys(table)) { if (key.length > best.length && id.startsWith(key) && /[-@:]/.test(id.charAt(key.length))) best = key; } if (!best) return {}; const entry = table[best]; if (!entry) return {}; const out = clone(entry); delete out.label; // the label names the base model, not this variant if (!SNAPSHOT_SUFFIX.test(id.slice(best.length))) { // A genuinely different variant (`o3-deep-research`, `gpt-4o-audio-...`): // capabilities are a fair guess, the price is not. delete out.priceIn; delete out.priceOut; delete out.priceCachedIn; } return out; } /** * Defaults for a model: exact id first, then the longest known id that prefixes * it (so dated snapshots match their base model). Returns `{}` when unknown. */ export function lookupModelDefaults(kind: ProviderKind, modelId: string): Partial { // Gemini's listing names models "models/". const id = modelId.trim().replace(/^models\//, ''); if (!id) return {}; if (kind === 'openai_compatible') { // Routers name models "/" (`anthropic/claude-sonnet-5-5`). // Borrow capabilities and limits from the vendor's table, never its prices. const bare = id.slice(id.lastIndexOf('/') + 1).replace(/:[a-z]+$/, ''); for (const table of [OPENAI, ANTHROPIC, GEMINI, DEEPSEEK]) { const hit = lookupIn(table, bare); if (Object.keys(hit).length > 0) { delete hit.priceIn; delete hit.priceOut; delete hit.priceCachedIn; // Temperature limits are properties of the first-party APIs, not of the router. if (hit.capabilities) delete hit.capabilities.temperature; return hit; } } return {}; } return lookupIn(TABLES[kind], id); } /** Every catalogued model id for a provider (used as a fallback list when discovery is unavailable). */ export function catalogModelIds(kind: ProviderKind): string[] { return Object.keys(TABLES[kind]); } /** * What a cache-hit input token costs when a model has no cached price of its own, as a * fraction of the normal input price. These are the providers' published discounts for * their current models: OpenAI GPT-5 and later, Claude cache reads and Gemini 2.5 and later * all bill cached input at 10%. Older OpenAI models (GPT-4.1 25%, GPT-4o / o-series 50%) * discount less. Servers behind the compatible adapter and Ollama have no known discount. */ export function cachedInputRatio(kind: ProviderKind, modelId: string): number { const id = modelId.slice(modelId.lastIndexOf('/') + 1); switch (kind) { case 'openai': if (/^gpt-4\.1/.test(id)) return 0.25; if (/^(gpt-4o|o\d)/.test(id)) return 0.5; return 0.1; case 'anthropic': case 'gemini': return 0.1; case 'deepseek': return 0.02; default: return 1; } } /** USD per million cached input tokens: the model's own price when set, otherwise the provider discount. */ export function cachedInputPrice(kind: ProviderKind, modelId: string, priceIn: number | null, priceCachedIn: number | null): number { if (priceCachedIn !== null && priceCachedIn !== undefined) return priceCachedIn; return (priceIn ?? 0) * cachedInputRatio(kind, modelId); } /** Cache writes: Claude bills writing the (5-minute) prompt cache at 1.25x the input price. */ export function cacheWriteRatio(kind: ProviderKind): number { return kind === 'anthropic' ? 1.25 : 1; } /** * Time-of-day pricing. DeepSeek charges half its list price outside peak hours * (01:00-04:00 and 06:00-10:00 UTC, Monday to Friday). */ export function priceMultiplier(kind: ProviderKind, at: number = Date.now()): number { if (kind !== 'deepseek') return 1; const d = new Date(at); const day = d.getUTCDay(); const hour = d.getUTCHours(); const peak = day >= 1 && day <= 5 && ((hour >= 1 && hour < 4) || (hour >= 6 && hour < 10)); return peak ? 1 : 0.5; }