mirror of
https://github.com/getcompanion-ai/co-mono.git
synced 2026-04-15 20:03:05 +00:00
refactor(ai): Update LLM implementations to use Model objects
- LLM constructors now take Model objects instead of string IDs - Added provider field to AssistantMessage interface - Updated getModel function with type-safe model ID autocomplete - Fixed Anthropic model ID mapping for proper API aliases - Added baseUrl to Model interface for provider-specific endpoints - Updated all tests to use getModel for model instantiation - Removed deprecated models.json in favor of generated models
This commit is contained in:
Mario Zechner
parent
d61d09b88d
commit
f9d688d577
11 changed files with 334 additions and 8447 deletions
|
|
@ -1,7 +1,12 @@
|
|||
#!/usr/bin/env tsx
|
||||
|
||||
import { readFileSync, writeFileSync } from "fs";
|
||||
import { join } from "path";
|
||||
import { join, dirname } from "path";
|
||||
import { fileURLToPath } from "url";
|
||||
|
||||
const __filename = fileURLToPath(import.meta.url);
|
||||
const __dirname = dirname(__filename);
|
||||
const packageRoot = join(__dirname, "..");
|
||||
|
||||
interface ModelsDevModel {
|
||||
id: string;
|
||||
|
|
@ -27,6 +32,7 @@ interface NormalizedModel {
|
|||
id: string;
|
||||
name: string;
|
||||
provider: string;
|
||||
baseUrl?: string;
|
||||
reasoning: boolean;
|
||||
input: ("text" | "image")[];
|
||||
cost: {
|
||||
|
|
@ -41,7 +47,7 @@ interface NormalizedModel {
|
|||
|
||||
async function fetchOpenRouterModels(): Promise<NormalizedModel[]> {
|
||||
try {
|
||||
console.log("🌐 Fetching models from OpenRouter API...");
|
||||
console.log("Fetching models from OpenRouter API...");
|
||||
const response = await fetch("https://openrouter.ai/api/v1/models");
|
||||
const data = await response.json();
|
||||
|
||||
|
|
@ -65,22 +71,39 @@ async function fetchOpenRouterModels(): Promise<NormalizedModel[]> {
|
|||
modelKey = model.id.replace("openai/", "");
|
||||
} else if (model.id.startsWith("anthropic/")) {
|
||||
provider = "anthropic";
|
||||
const fullKey = model.id.replace("anthropic/", "");
|
||||
// Map to Anthropic's preferred aliases
|
||||
const anthropicAliases: Record<string, string> = {
|
||||
"claude-opus-4.1": "claude-opus-4-1",
|
||||
"claude-opus-4": "claude-opus-4-0",
|
||||
"claude-sonnet-4": "claude-sonnet-4-0",
|
||||
"claude-3.7-sonnet": "claude-3-7-sonnet-latest",
|
||||
"claude-3.7-sonnet:thinking": "claude-3-7-sonnet-latest:thinking",
|
||||
"claude-3.5-haiku": "claude-3-5-haiku-latest",
|
||||
"claude-3.5-haiku-20241022": "claude-3-5-haiku-latest",
|
||||
"claude-3-haiku": "claude-3-haiku-20240307",
|
||||
"claude-3-sonnet": "claude-3-sonnet-20240229",
|
||||
"claude-3-opus": "claude-3-opus-20240229",
|
||||
"claude-3.5-sonnet": "claude-3-5-sonnet-latest"
|
||||
};
|
||||
modelKey = anthropicAliases[fullKey] || fullKey;
|
||||
modelKey = model.id.replace("anthropic/", "");
|
||||
|
||||
// Fix dot notation to dash notation for ALL Anthropic models
|
||||
modelKey = modelKey.replace(/\./g, "-");
|
||||
|
||||
// Map version-less models to -latest aliases
|
||||
if (modelKey === "claude-3-5-haiku") {
|
||||
modelKey = "claude-3-5-haiku-latest";
|
||||
} else if (modelKey === "claude-3-5-sonnet") {
|
||||
modelKey = "claude-3-5-sonnet-latest";
|
||||
} else if (modelKey === "claude-3-7-sonnet") {
|
||||
modelKey = "claude-3-7-sonnet-latest";
|
||||
} else if (modelKey === "claude-3-7-sonnet:thinking") {
|
||||
modelKey = "claude-3-7-sonnet-latest:thinking";
|
||||
}
|
||||
// Map numbered versions to proper format
|
||||
else if (modelKey === "claude-opus-4-1") {
|
||||
modelKey = "claude-opus-4-1";
|
||||
} else if (modelKey === "claude-opus-4") {
|
||||
modelKey = "claude-opus-4-0";
|
||||
} else if (modelKey === "claude-sonnet-4") {
|
||||
modelKey = "claude-sonnet-4-0";
|
||||
}
|
||||
// Map old 3.x models to their specific dates
|
||||
else if (modelKey === "claude-3-haiku") {
|
||||
modelKey = "claude-3-haiku-20240307";
|
||||
} else if (modelKey === "claude-3-sonnet") {
|
||||
modelKey = "claude-3-sonnet-20240229";
|
||||
} else if (modelKey === "claude-3-opus") {
|
||||
modelKey = "claude-3-opus-20240229";
|
||||
} else {
|
||||
modelKey = modelKey.replace("\.", "-");
|
||||
}
|
||||
} else if (model.id.startsWith("x-ai/")) {
|
||||
provider = "xai";
|
||||
modelKey = model.id.replace("x-ai/", "");
|
||||
|
|
@ -107,7 +130,7 @@ async function fetchOpenRouterModels(): Promise<NormalizedModel[]> {
|
|||
const cacheReadCost = parseFloat(model.pricing?.input_cache_read || "0") * 1_000_000;
|
||||
const cacheWriteCost = parseFloat(model.pricing?.input_cache_write || "0") * 1_000_000;
|
||||
|
||||
models.push({
|
||||
const normalizedModel: NormalizedModel = {
|
||||
id: modelKey,
|
||||
name: model.name,
|
||||
provider,
|
||||
|
|
@ -121,21 +144,30 @@ async function fetchOpenRouterModels(): Promise<NormalizedModel[]> {
|
|||
},
|
||||
contextWindow: model.context_length || 4096,
|
||||
maxTokens: model.top_provider?.max_completion_tokens || 4096,
|
||||
});
|
||||
};
|
||||
|
||||
// Add baseUrl for providers that need it
|
||||
if (provider === "xai") {
|
||||
normalizedModel.baseUrl = "https://api.x.ai/v1";
|
||||
} else if (provider === "openrouter") {
|
||||
normalizedModel.baseUrl = "https://openrouter.ai/api/v1";
|
||||
}
|
||||
|
||||
models.push(normalizedModel);
|
||||
}
|
||||
|
||||
console.log(`✅ Fetched ${models.length} tool-capable models from OpenRouter`);
|
||||
console.log(`Fetched ${models.length} tool-capable models from OpenRouter`);
|
||||
return models;
|
||||
} catch (error) {
|
||||
console.error("❌ Failed to fetch OpenRouter models:", error);
|
||||
console.error("Failed to fetch OpenRouter models:", error);
|
||||
return [];
|
||||
}
|
||||
}
|
||||
|
||||
function loadModelsDevData(): NormalizedModel[] {
|
||||
try {
|
||||
console.log("📁 Loading models from models.json...");
|
||||
const data = JSON.parse(readFileSync(join(process.cwd(), "src/models.json"), "utf-8"));
|
||||
console.log("Loading models from models.json...");
|
||||
const data = JSON.parse(readFileSync(join(packageRoot, "src/models.json"), "utf-8"));
|
||||
|
||||
const models: NormalizedModel[] = [];
|
||||
|
||||
|
|
@ -149,6 +181,7 @@ function loadModelsDevData(): NormalizedModel[] {
|
|||
id: modelId,
|
||||
name: m.name || modelId,
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: m.reasoning === true,
|
||||
input: m.modalities?.input?.includes("image") ? ["text", "image"] : ["text"],
|
||||
cost: {
|
||||
|
|
@ -173,6 +206,7 @@ function loadModelsDevData(): NormalizedModel[] {
|
|||
id: modelId,
|
||||
name: m.name || modelId,
|
||||
provider: "cerebras",
|
||||
baseUrl: "https://api.cerebras.ai/v1",
|
||||
reasoning: m.reasoning === true,
|
||||
input: m.modalities?.input?.includes("image") ? ["text", "image"] : ["text"],
|
||||
cost: {
|
||||
|
|
@ -187,10 +221,10 @@ function loadModelsDevData(): NormalizedModel[] {
|
|||
}
|
||||
}
|
||||
|
||||
console.log(`✅ Loaded ${models.length} tool-capable models from models.dev`);
|
||||
console.log(`Loaded ${models.length} tool-capable models from models.dev`);
|
||||
return models;
|
||||
} catch (error) {
|
||||
console.error("❌ Failed to load models.dev data:", error);
|
||||
console.error("Failed to load models.dev data:", error);
|
||||
return [];
|
||||
}
|
||||
}
|
||||
|
|
@ -235,6 +269,9 @@ export const PROVIDERS = {
|
|||
output += `\t\t\t\tid: "${model.id}",\n`;
|
||||
output += `\t\t\t\tname: "${model.name}",\n`;
|
||||
output += `\t\t\t\tprovider: "${model.provider}",\n`;
|
||||
if (model.baseUrl) {
|
||||
output += `\t\t\t\tbaseUrl: "${model.baseUrl}",\n`;
|
||||
}
|
||||
output += `\t\t\t\treasoning: ${model.reasoning},\n`;
|
||||
output += `\t\t\t\tinput: ${JSON.stringify(model.input)},\n`;
|
||||
output += `\t\t\t\tcost: {\n`;
|
||||
|
|
@ -261,14 +298,14 @@ export type ProviderModels = {
|
|||
`;
|
||||
|
||||
// Write file
|
||||
writeFileSync(join(process.cwd(), "src/models.generated.ts"), output);
|
||||
console.log("✅ Generated src/models.generated.ts");
|
||||
writeFileSync(join(packageRoot, "src/models.generated.ts"), output);
|
||||
console.log("Generated src/models.generated.ts");
|
||||
|
||||
// Print statistics
|
||||
const totalModels = allModels.length;
|
||||
const reasoningModels = allModels.filter(m => m.reasoning).length;
|
||||
|
||||
console.log(`\n📊 Model Statistics:`);
|
||||
console.log(`\nModel Statistics:`);
|
||||
console.log(` Total tool-capable models: ${totalModels}`);
|
||||
console.log(` Reasoning-capable models: ${reasoningModels}`);
|
||||
|
||||
|
|
|
|||
|
|
@ -24,7 +24,7 @@ export {
|
|||
|
||||
// Export providers
|
||||
export { AnthropicLLM } from "./providers/anthropic.js";
|
||||
export { GoogleLLM } from "./providers/gemini.js";
|
||||
export { GoogleLLM } from "./providers/google.js";
|
||||
export { OpenAICompletionsLLM } from "./providers/openai-completions.js";
|
||||
export { OpenAIResponsesLLM } from "./providers/openai-responses.js";
|
||||
|
||||
|
|
|
|||
|
|
@ -10,6 +10,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek-r1-distill-llama-70b",
|
||||
name: "DeepSeek R1 Distill Llama 70B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -25,6 +26,7 @@ export const PROVIDERS = {
|
|||
id: "llama3-70b-8192",
|
||||
name: "Llama 3 70B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -40,6 +42,7 @@ export const PROVIDERS = {
|
|||
id: "llama-3.3-70b-versatile",
|
||||
name: "Llama 3.3 70B Versatile",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -55,6 +58,7 @@ export const PROVIDERS = {
|
|||
id: "llama-3.1-8b-instant",
|
||||
name: "Llama 3.1 8B Instant",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -70,6 +74,7 @@ export const PROVIDERS = {
|
|||
id: "qwen-qwq-32b",
|
||||
name: "Qwen QwQ 32B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -85,6 +90,7 @@ export const PROVIDERS = {
|
|||
id: "gemma2-9b-it",
|
||||
name: "Gemma 2 9B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -100,6 +106,7 @@ export const PROVIDERS = {
|
|||
id: "mistral-saba-24b",
|
||||
name: "Mistral Saba 24B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -115,6 +122,7 @@ export const PROVIDERS = {
|
|||
id: "llama3-8b-8192",
|
||||
name: "Llama 3 8B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -130,6 +138,7 @@ export const PROVIDERS = {
|
|||
id: "openai/gpt-oss-120b",
|
||||
name: "GPT OSS 120B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -145,6 +154,7 @@ export const PROVIDERS = {
|
|||
id: "openai/gpt-oss-20b",
|
||||
name: "GPT OSS 20B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -160,6 +170,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-4-maverick-17b-128e-instruct",
|
||||
name: "Llama 4 Maverick 17B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -175,6 +186,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-4-scout-17b-16e-instruct",
|
||||
name: "Llama 4 Scout 17B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -190,6 +202,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-32b",
|
||||
name: "Qwen3 32B",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -205,6 +218,7 @@ export const PROVIDERS = {
|
|||
id: "moonshotai/kimi-k2-instruct",
|
||||
name: "Kimi K2 Instruct",
|
||||
provider: "groq",
|
||||
baseUrl: "https://api.groq.com/openai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -224,6 +238,7 @@ export const PROVIDERS = {
|
|||
id: "qwen-3-235b-a22b-instruct-2507",
|
||||
name: "Qwen 3 235B Instruct",
|
||||
provider: "cerebras",
|
||||
baseUrl: "https://api.cerebras.ai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -239,6 +254,7 @@ export const PROVIDERS = {
|
|||
id: "gpt-oss-120b",
|
||||
name: "GPT OSS 120B",
|
||||
provider: "cerebras",
|
||||
baseUrl: "https://api.cerebras.ai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -254,6 +270,7 @@ export const PROVIDERS = {
|
|||
id: "qwen-3-coder-480b",
|
||||
name: "Qwen 3 Coder 480B",
|
||||
provider: "cerebras",
|
||||
baseUrl: "https://api.cerebras.ai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -273,6 +290,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-30b-a3b-thinking-2507",
|
||||
name: "Qwen: Qwen3 30B A3B Thinking 2507",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -288,6 +306,7 @@ export const PROVIDERS = {
|
|||
id: "nousresearch/hermes-4-70b",
|
||||
name: "Nous: Hermes 4 70B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -303,6 +322,7 @@ export const PROVIDERS = {
|
|||
id: "nousresearch/hermes-4-405b",
|
||||
name: "Nous: Hermes 4 405B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -318,6 +338,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-chat-v3.1:free",
|
||||
name: "DeepSeek: DeepSeek V3.1 (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -333,6 +354,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-chat-v3.1",
|
||||
name: "DeepSeek: DeepSeek V3.1",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -348,6 +370,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-medium-3.1",
|
||||
name: "Mistral: Mistral Medium 3.1",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -363,6 +386,7 @@ export const PROVIDERS = {
|
|||
id: "z-ai/glm-4.5v",
|
||||
name: "Z.AI: GLM 4.5V",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -378,6 +402,7 @@ export const PROVIDERS = {
|
|||
id: "ai21/jamba-mini-1.7",
|
||||
name: "AI21: Jamba Mini 1.7",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -393,6 +418,7 @@ export const PROVIDERS = {
|
|||
id: "ai21/jamba-large-1.7",
|
||||
name: "AI21: Jamba Large 1.7",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -408,6 +434,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/codestral-2508",
|
||||
name: "Mistral: Codestral 2508",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -423,6 +450,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-coder-30b-a3b-instruct",
|
||||
name: "Qwen: Qwen3 Coder 30B A3B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -438,6 +466,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-30b-a3b-instruct-2507",
|
||||
name: "Qwen: Qwen3 30B A3B Instruct 2507",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -453,6 +482,7 @@ export const PROVIDERS = {
|
|||
id: "z-ai/glm-4.5",
|
||||
name: "Z.AI: GLM 4.5",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -468,6 +498,7 @@ export const PROVIDERS = {
|
|||
id: "z-ai/glm-4.5-air:free",
|
||||
name: "Z.AI: GLM 4.5 Air (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -483,6 +514,7 @@ export const PROVIDERS = {
|
|||
id: "z-ai/glm-4.5-air",
|
||||
name: "Z.AI: GLM 4.5 Air",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -498,6 +530,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-235b-a22b-thinking-2507",
|
||||
name: "Qwen: Qwen3 235B A22B Thinking 2507",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -513,6 +546,7 @@ export const PROVIDERS = {
|
|||
id: "z-ai/glm-4-32b",
|
||||
name: "Z.AI: GLM 4 32B ",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -528,6 +562,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-coder:free",
|
||||
name: "Qwen: Qwen3 Coder 480B A35B (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -543,6 +578,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-coder",
|
||||
name: "Qwen: Qwen3 Coder 480B A35B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -558,6 +594,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-235b-a22b-2507",
|
||||
name: "Qwen: Qwen3 235B A22B Instruct 2507",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -573,6 +610,7 @@ export const PROVIDERS = {
|
|||
id: "moonshotai/kimi-k2:free",
|
||||
name: "MoonshotAI: Kimi K2 (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -588,6 +626,7 @@ export const PROVIDERS = {
|
|||
id: "moonshotai/kimi-k2",
|
||||
name: "MoonshotAI: Kimi K2",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -603,6 +642,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/devstral-medium",
|
||||
name: "Mistral: Devstral Medium",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -618,6 +658,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/devstral-small",
|
||||
name: "Mistral: Devstral Small 1.1",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -633,6 +674,7 @@ export const PROVIDERS = {
|
|||
id: "inception/mercury",
|
||||
name: "Inception: Mercury",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -648,6 +690,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-small-3.2-24b-instruct:free",
|
||||
name: "Mistral: Mistral Small 3.2 24B (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -663,6 +706,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-small-3.2-24b-instruct",
|
||||
name: "Mistral: Mistral Small 3.2 24B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -678,6 +722,7 @@ export const PROVIDERS = {
|
|||
id: "minimax/minimax-m1",
|
||||
name: "MiniMax: MiniMax M1",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -693,6 +738,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/magistral-small-2506",
|
||||
name: "Mistral: Magistral Small 2506",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -708,6 +754,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/magistral-medium-2506",
|
||||
name: "Mistral: Magistral Medium 2506",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -723,6 +770,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/magistral-medium-2506:thinking",
|
||||
name: "Mistral: Magistral Medium 2506 (thinking)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -738,6 +786,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-r1-0528",
|
||||
name: "DeepSeek: R1 0528",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -753,6 +802,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/devstral-small-2505:free",
|
||||
name: "Mistral: Devstral Small 2505 (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -768,6 +818,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/devstral-small-2505",
|
||||
name: "Mistral: Devstral Small 2505",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -783,6 +834,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.3-8b-instruct:free",
|
||||
name: "Meta: Llama 3.3 8B Instruct (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -798,6 +850,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-medium-3",
|
||||
name: "Mistral: Mistral Medium 3",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -813,6 +866,7 @@ export const PROVIDERS = {
|
|||
id: "arcee-ai/virtuoso-large",
|
||||
name: "Arcee AI: Virtuoso Large",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -828,6 +882,7 @@ export const PROVIDERS = {
|
|||
id: "inception/mercury-coder",
|
||||
name: "Inception: Mercury Coder",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -843,6 +898,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-4b:free",
|
||||
name: "Qwen: Qwen3 4B (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -858,6 +914,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-30b-a3b",
|
||||
name: "Qwen: Qwen3 30B A3B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -873,6 +930,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-14b",
|
||||
name: "Qwen: Qwen3 14B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -888,6 +946,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-32b",
|
||||
name: "Qwen: Qwen3 32B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -903,6 +962,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-235b-a22b:free",
|
||||
name: "Qwen: Qwen3 235B A22B (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -918,6 +978,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen3-235b-a22b",
|
||||
name: "Qwen: Qwen3 235B A22B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -933,6 +994,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-4-maverick:free",
|
||||
name: "Meta: Llama 4 Maverick (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -948,6 +1010,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-4-maverick",
|
||||
name: "Meta: Llama 4 Maverick",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -963,6 +1026,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-4-scout:free",
|
||||
name: "Meta: Llama 4 Scout (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -978,6 +1042,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-4-scout",
|
||||
name: "Meta: Llama 4 Scout",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -993,6 +1058,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-chat-v3-0324:free",
|
||||
name: "DeepSeek: DeepSeek V3 0324 (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1008,6 +1074,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-chat-v3-0324",
|
||||
name: "DeepSeek: DeepSeek V3 0324",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1023,6 +1090,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-small-3.1-24b-instruct:free",
|
||||
name: "Mistral: Mistral Small 3.1 24B (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1038,6 +1106,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-small-3.1-24b-instruct",
|
||||
name: "Mistral: Mistral Small 3.1 24B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1053,6 +1122,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-saba",
|
||||
name: "Mistral: Saba",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1068,6 +1138,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen-turbo",
|
||||
name: "Qwen: Qwen-Turbo",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1083,6 +1154,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen-plus",
|
||||
name: "Qwen: Qwen-Plus",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1098,6 +1170,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen-max",
|
||||
name: "Qwen: Qwen-Max ",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1113,6 +1186,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-small-24b-instruct-2501",
|
||||
name: "Mistral: Mistral Small 3",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1128,6 +1202,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-r1-distill-llama-70b",
|
||||
name: "DeepSeek: R1 Distill Llama 70B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1143,6 +1218,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-r1",
|
||||
name: "DeepSeek: R1",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1158,6 +1234,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/codestral-2501",
|
||||
name: "Mistral: Codestral 2501",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1173,6 +1250,7 @@ export const PROVIDERS = {
|
|||
id: "deepseek/deepseek-chat",
|
||||
name: "DeepSeek: DeepSeek V3",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1188,6 +1266,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.3-70b-instruct:free",
|
||||
name: "Meta: Llama 3.3 70B Instruct (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1203,6 +1282,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.3-70b-instruct",
|
||||
name: "Meta: Llama 3.3 70B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1218,6 +1298,7 @@ export const PROVIDERS = {
|
|||
id: "amazon/nova-lite-v1",
|
||||
name: "Amazon: Nova Lite 1.0",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1233,6 +1314,7 @@ export const PROVIDERS = {
|
|||
id: "amazon/nova-micro-v1",
|
||||
name: "Amazon: Nova Micro 1.0",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1248,6 +1330,7 @@ export const PROVIDERS = {
|
|||
id: "amazon/nova-pro-v1",
|
||||
name: "Amazon: Nova Pro 1.0",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1263,6 +1346,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-large-2411",
|
||||
name: "Mistral Large 2411",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1278,6 +1362,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-large-2407",
|
||||
name: "Mistral Large 2407",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1293,6 +1378,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/pixtral-large-2411",
|
||||
name: "Mistral: Pixtral Large 2411",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1308,6 +1394,7 @@ export const PROVIDERS = {
|
|||
id: "thedrummer/unslopnemo-12b",
|
||||
name: "TheDrummer: UnslopNemo 12B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1323,6 +1410,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/ministral-8b",
|
||||
name: "Mistral: Ministral 8B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1338,6 +1426,7 @@ export const PROVIDERS = {
|
|||
id: "nvidia/llama-3.1-nemotron-70b-instruct",
|
||||
name: "NVIDIA: Llama 3.1 Nemotron 70B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1353,6 +1442,7 @@ export const PROVIDERS = {
|
|||
id: "thedrummer/rocinante-12b",
|
||||
name: "TheDrummer: Rocinante 12B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1368,6 +1458,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.2-3b-instruct",
|
||||
name: "Meta: Llama 3.2 3B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1383,6 +1474,7 @@ export const PROVIDERS = {
|
|||
id: "qwen/qwen-2.5-72b-instruct",
|
||||
name: "Qwen2.5 72B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1398,6 +1490,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/pixtral-12b",
|
||||
name: "Mistral: Pixtral 12B",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1413,6 +1506,7 @@ export const PROVIDERS = {
|
|||
id: "cohere/command-r-plus-08-2024",
|
||||
name: "Cohere: Command R+ (08-2024)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1428,6 +1522,7 @@ export const PROVIDERS = {
|
|||
id: "cohere/command-r-08-2024",
|
||||
name: "Cohere: Command R (08-2024)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1443,6 +1538,7 @@ export const PROVIDERS = {
|
|||
id: "microsoft/phi-3.5-mini-128k-instruct",
|
||||
name: "Microsoft: Phi-3.5 Mini 128K Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1458,6 +1554,7 @@ export const PROVIDERS = {
|
|||
id: "nousresearch/hermes-3-llama-3.1-70b",
|
||||
name: "Nous: Hermes 3 70B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1473,6 +1570,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.1-8b-instruct",
|
||||
name: "Meta: Llama 3.1 8B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1488,6 +1586,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.1-70b-instruct",
|
||||
name: "Meta: Llama 3.1 70B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1503,6 +1602,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3.1-405b-instruct",
|
||||
name: "Meta: Llama 3.1 405B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1518,6 +1618,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-nemo",
|
||||
name: "Mistral: Mistral Nemo",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1533,6 +1634,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-7b-instruct-v0.3",
|
||||
name: "Mistral: Mistral 7B Instruct v0.3",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1548,6 +1650,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-7b-instruct:free",
|
||||
name: "Mistral: Mistral 7B Instruct (free)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1563,6 +1666,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-7b-instruct",
|
||||
name: "Mistral: Mistral 7B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1578,6 +1682,7 @@ export const PROVIDERS = {
|
|||
id: "microsoft/phi-3-mini-128k-instruct",
|
||||
name: "Microsoft: Phi-3 Mini 128K Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1593,6 +1698,7 @@ export const PROVIDERS = {
|
|||
id: "microsoft/phi-3-medium-128k-instruct",
|
||||
name: "Microsoft: Phi-3 Medium 128K Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1608,6 +1714,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3-70b-instruct",
|
||||
name: "Meta: Llama 3 70B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1623,6 +1730,7 @@ export const PROVIDERS = {
|
|||
id: "meta-llama/llama-3-8b-instruct",
|
||||
name: "Meta: Llama 3 8B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1638,6 +1746,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mixtral-8x22b-instruct",
|
||||
name: "Mistral: Mixtral 8x22B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1653,6 +1762,7 @@ export const PROVIDERS = {
|
|||
id: "cohere/command-r-plus",
|
||||
name: "Cohere: Command R+",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1668,6 +1778,7 @@ export const PROVIDERS = {
|
|||
id: "cohere/command-r-plus-04-2024",
|
||||
name: "Cohere: Command R+ (04-2024)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1683,6 +1794,7 @@ export const PROVIDERS = {
|
|||
id: "cohere/command-r",
|
||||
name: "Cohere: Command R",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1698,6 +1810,7 @@ export const PROVIDERS = {
|
|||
id: "cohere/command-r-03-2024",
|
||||
name: "Cohere: Command R (03-2024)",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1713,6 +1826,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-large",
|
||||
name: "Mistral Large",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1728,6 +1842,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-tiny",
|
||||
name: "Mistral Tiny",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1743,6 +1858,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-small",
|
||||
name: "Mistral Small",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1758,6 +1874,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mixtral-8x7b-instruct",
|
||||
name: "Mistral: Mixtral 8x7B Instruct",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1773,6 +1890,7 @@ export const PROVIDERS = {
|
|||
id: "mistralai/mistral-7b-instruct-v0.1",
|
||||
name: "Mistral: Mistral 7B Instruct v0.1",
|
||||
provider: "openrouter",
|
||||
baseUrl: "https://openrouter.ai/api/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1792,6 +1910,7 @@ export const PROVIDERS = {
|
|||
id: "grok-code-fast-1",
|
||||
name: "xAI: Grok Code Fast 1",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1807,6 +1926,7 @@ export const PROVIDERS = {
|
|||
id: "grok-4",
|
||||
name: "xAI: Grok 4",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: true,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
|
|
@ -1822,6 +1942,7 @@ export const PROVIDERS = {
|
|||
id: "grok-3-mini",
|
||||
name: "xAI: Grok 3 Mini",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1837,6 +1958,7 @@ export const PROVIDERS = {
|
|||
id: "grok-3",
|
||||
name: "xAI: Grok 3",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1852,6 +1974,7 @@ export const PROVIDERS = {
|
|||
id: "grok-3-mini-beta",
|
||||
name: "xAI: Grok 3 Mini Beta",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1867,6 +1990,7 @@ export const PROVIDERS = {
|
|||
id: "grok-3-beta",
|
||||
name: "xAI: Grok 3 Beta",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -1882,6 +2006,7 @@ export const PROVIDERS = {
|
|||
id: "grok-2-1212",
|
||||
name: "xAI: Grok 2 1212",
|
||||
provider: "xai",
|
||||
baseUrl: "https://api.x.ai/v1",
|
||||
reasoning: false,
|
||||
input: ["text"],
|
||||
cost: {
|
||||
|
|
@ -2456,9 +2581,24 @@ export const PROVIDERS = {
|
|||
contextWindow: 200000,
|
||||
maxTokens: 64000,
|
||||
} satisfies Model,
|
||||
"claude-3-5-haiku-20241022": {
|
||||
id: "claude-3-5-haiku-20241022",
|
||||
name: "Anthropic: Claude 3.5 Haiku (2024-10-22)",
|
||||
provider: "anthropic",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
cost: {
|
||||
input: 0.7999999999999999,
|
||||
output: 4,
|
||||
cacheRead: 0.08,
|
||||
cacheWrite: 1,
|
||||
},
|
||||
contextWindow: 200000,
|
||||
maxTokens: 8192,
|
||||
} satisfies Model,
|
||||
"claude-3-5-haiku-latest": {
|
||||
id: "claude-3-5-haiku-latest",
|
||||
name: "Anthropic: Claude 3.5 Haiku (2024-10-22)",
|
||||
name: "Anthropic: Claude 3.5 Haiku",
|
||||
provider: "anthropic",
|
||||
reasoning: false,
|
||||
input: ["text", "image"],
|
||||
|
|
@ -2486,8 +2626,8 @@ export const PROVIDERS = {
|
|||
contextWindow: 200000,
|
||||
maxTokens: 8192,
|
||||
} satisfies Model,
|
||||
"claude-3.5-sonnet-20240620": {
|
||||
id: "claude-3.5-sonnet-20240620",
|
||||
"claude-3-5-sonnet-20240620": {
|
||||
id: "claude-3-5-sonnet-20240620",
|
||||
name: "Anthropic: Claude 3.5 Sonnet (2024-06-20)",
|
||||
provider: "anthropic",
|
||||
reasoning: false,
|
||||
|
|
|
|||
File diff suppressed because it is too large
Load diff
|
|
@ -1,6 +1,6 @@
|
|||
import { PROVIDERS } from "./models.generated.js";
|
||||
import { AnthropicLLM } from "./providers/anthropic.js";
|
||||
import { GoogleLLM } from "./providers/gemini.js";
|
||||
import { GoogleLLM } from "./providers/google.js";
|
||||
import { OpenAICompletionsLLM } from "./providers/openai-completions.js";
|
||||
import { OpenAIResponsesLLM } from "./providers/openai-responses.js";
|
||||
import type { Model } from "./types.js";
|
||||
|
|
@ -9,33 +9,31 @@ import type { Model } from "./types.js";
|
|||
export const PROVIDER_CONFIG = {
|
||||
google: {
|
||||
envKey: "GEMINI_API_KEY",
|
||||
create: (model: string, apiKey: string) => new GoogleLLM(model, apiKey),
|
||||
create: (model: Model, apiKey: string) => new GoogleLLM(model, apiKey),
|
||||
},
|
||||
openai: {
|
||||
envKey: "OPENAI_API_KEY",
|
||||
create: (model: string, apiKey: string) => new OpenAIResponsesLLM(model, apiKey),
|
||||
create: (model: Model, apiKey: string) => new OpenAIResponsesLLM(model, apiKey),
|
||||
},
|
||||
anthropic: {
|
||||
envKey: "ANTHROPIC_API_KEY",
|
||||
create: (model: string, apiKey: string) => new AnthropicLLM(model, apiKey),
|
||||
create: (model: Model, apiKey: string) => new AnthropicLLM(model, apiKey),
|
||||
},
|
||||
xai: {
|
||||
envKey: "XAI_API_KEY",
|
||||
create: (model: string, apiKey: string) => new OpenAICompletionsLLM(model, apiKey, "https://api.x.ai/v1"),
|
||||
create: (model: Model, apiKey: string) => new OpenAICompletionsLLM(model, apiKey),
|
||||
},
|
||||
groq: {
|
||||
envKey: "GROQ_API_KEY",
|
||||
create: (model: string, apiKey: string) =>
|
||||
new OpenAICompletionsLLM(model, apiKey, "https://api.groq.com/openai/v1"),
|
||||
create: (model: Model, apiKey: string) => new OpenAICompletionsLLM(model, apiKey),
|
||||
},
|
||||
cerebras: {
|
||||
envKey: "CEREBRAS_API_KEY",
|
||||
create: (model: string, apiKey: string) => new OpenAICompletionsLLM(model, apiKey, "https://api.cerebras.ai/v1"),
|
||||
create: (model: Model, apiKey: string) => new OpenAICompletionsLLM(model, apiKey),
|
||||
},
|
||||
openrouter: {
|
||||
envKey: "OPENROUTER_API_KEY",
|
||||
create: (model: string, apiKey: string) =>
|
||||
new OpenAICompletionsLLM(model, apiKey, "https://openrouter.ai/api/v1"),
|
||||
create: (model: Model, apiKey: string) => new OpenAICompletionsLLM(model, apiKey),
|
||||
},
|
||||
} as const;
|
||||
|
||||
|
|
@ -90,7 +88,18 @@ export function createLLM<P extends keyof typeof PROVIDERS, M extends keyof (typ
|
|||
const key = apiKey || process.env[config.envKey];
|
||||
if (!key) throw new Error(`No API key provided for ${provider}. Set ${config.envKey} or pass apiKey.`);
|
||||
|
||||
return config.create(model as string, key) as ProviderToLLM[P];
|
||||
return config.create(modelData, key) as ProviderToLLM[P];
|
||||
}
|
||||
|
||||
// Helper function to get model info with type-safe model IDs
|
||||
export function getModel<P extends keyof typeof PROVIDERS>(
|
||||
provider: P,
|
||||
modelId: keyof (typeof PROVIDERS)[P]["models"],
|
||||
): Model | undefined {
|
||||
const providerData = PROVIDERS[provider];
|
||||
if (!providerData) return undefined;
|
||||
const models = providerData.models as Record<string, Model>;
|
||||
return models[modelId as string];
|
||||
}
|
||||
|
||||
// Re-export Model type for convenience
|
||||
|
|
|
|||
|
|
@ -11,6 +11,7 @@ import type {
|
|||
LLM,
|
||||
LLMOptions,
|
||||
Message,
|
||||
Model,
|
||||
StopReason,
|
||||
TokenUsage,
|
||||
ToolCall,
|
||||
|
|
@ -26,10 +27,10 @@ export interface AnthropicLLMOptions extends LLMOptions {
|
|||
|
||||
export class AnthropicLLM implements LLM<AnthropicLLMOptions> {
|
||||
private client: Anthropic;
|
||||
private model: string;
|
||||
private modelInfo: Model;
|
||||
private isOAuthToken: boolean = false;
|
||||
|
||||
constructor(model: string, apiKey?: string, baseUrl?: string) {
|
||||
constructor(model: Model, apiKey?: string) {
|
||||
if (!apiKey) {
|
||||
if (!process.env.ANTHROPIC_API_KEY) {
|
||||
throw new Error(
|
||||
|
|
@ -45,13 +46,17 @@ export class AnthropicLLM implements LLM<AnthropicLLMOptions> {
|
|||
};
|
||||
|
||||
process.env.ANTHROPIC_API_KEY = undefined;
|
||||
this.client = new Anthropic({ apiKey: null, authToken: apiKey, baseURL: baseUrl, defaultHeaders });
|
||||
this.client = new Anthropic({ apiKey: null, authToken: apiKey, baseURL: model.baseUrl, defaultHeaders });
|
||||
this.isOAuthToken = true;
|
||||
} else {
|
||||
this.client = new Anthropic({ apiKey, baseURL: baseUrl });
|
||||
this.client = new Anthropic({ apiKey, baseURL: model.baseUrl });
|
||||
this.isOAuthToken = false;
|
||||
}
|
||||
this.model = model;
|
||||
this.modelInfo = model;
|
||||
}
|
||||
|
||||
getModel(): Model {
|
||||
return this.modelInfo;
|
||||
}
|
||||
|
||||
async complete(context: Context, options?: AnthropicLLMOptions): Promise<AssistantMessage> {
|
||||
|
|
@ -59,7 +64,7 @@ export class AnthropicLLM implements LLM<AnthropicLLMOptions> {
|
|||
const messages = this.convertMessages(context.messages);
|
||||
|
||||
const params: MessageCreateParamsStreaming = {
|
||||
model: this.model,
|
||||
model: this.modelInfo.id,
|
||||
messages,
|
||||
max_tokens: options?.maxTokens || 4096,
|
||||
stream: true,
|
||||
|
|
@ -97,7 +102,8 @@ export class AnthropicLLM implements LLM<AnthropicLLMOptions> {
|
|||
params.tools = this.convertTools(context.tools);
|
||||
}
|
||||
|
||||
if (options?.thinking?.enabled) {
|
||||
// Only enable thinking if the model supports it
|
||||
if (options?.thinking?.enabled && this.modelInfo.reasoning) {
|
||||
params.thinking = {
|
||||
type: "enabled",
|
||||
budget_tokens: options.thinking.budgetTokens || 1024,
|
||||
|
|
@ -194,14 +200,16 @@ export class AnthropicLLM implements LLM<AnthropicLLMOptions> {
|
|||
thinking,
|
||||
thinkingSignature,
|
||||
toolCalls,
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage,
|
||||
stopReason: this.mapStopReason(msg.stop_reason),
|
||||
};
|
||||
} catch (error) {
|
||||
return {
|
||||
role: "assistant",
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage: {
|
||||
input: 0,
|
||||
output: 0,
|
||||
|
|
|
|||
|
|
@ -11,6 +11,7 @@ import type {
|
|||
LLM,
|
||||
LLMOptions,
|
||||
Message,
|
||||
Model,
|
||||
StopReason,
|
||||
TokenUsage,
|
||||
Tool,
|
||||
|
|
@ -27,9 +28,9 @@ export interface GoogleLLMOptions extends LLMOptions {
|
|||
|
||||
export class GoogleLLM implements LLM<GoogleLLMOptions> {
|
||||
private client: GoogleGenAI;
|
||||
private model: string;
|
||||
private model: Model;
|
||||
|
||||
constructor(model: string, apiKey?: string) {
|
||||
constructor(model: Model, apiKey?: string) {
|
||||
if (!apiKey) {
|
||||
if (!process.env.GEMINI_API_KEY) {
|
||||
throw new Error(
|
||||
|
|
@ -42,6 +43,10 @@ export class GoogleLLM implements LLM<GoogleLLMOptions> {
|
|||
this.model = model;
|
||||
}
|
||||
|
||||
getModel(): Model {
|
||||
return this.model;
|
||||
}
|
||||
|
||||
async complete(context: Context, options?: GoogleLLMOptions): Promise<AssistantMessage> {
|
||||
try {
|
||||
const contents = this.convertMessages(context.messages);
|
||||
|
|
@ -71,8 +76,8 @@ export class GoogleLLM implements LLM<GoogleLLMOptions> {
|
|||
};
|
||||
}
|
||||
|
||||
// Add thinking config if enabled
|
||||
if (options?.thinking?.enabled) {
|
||||
// Add thinking config if enabled and model supports it
|
||||
if (options?.thinking?.enabled && this.model.reasoning) {
|
||||
config.thinkingConfig = {
|
||||
includeThoughts: true,
|
||||
...(options.thinking.budgetTokens !== undefined && { thinkingBudget: options.thinking.budgetTokens }),
|
||||
|
|
@ -81,7 +86,7 @@ export class GoogleLLM implements LLM<GoogleLLMOptions> {
|
|||
|
||||
// Build the request parameters
|
||||
const params: GenerateContentParameters = {
|
||||
model: this.model,
|
||||
model: this.model.id,
|
||||
contents,
|
||||
config,
|
||||
};
|
||||
|
|
@ -207,14 +212,16 @@ export class GoogleLLM implements LLM<GoogleLLMOptions> {
|
|||
thinking: thinking || undefined,
|
||||
thinkingSignature: thoughtSignature,
|
||||
toolCalls: toolCalls.length > 0 ? toolCalls : undefined,
|
||||
model: this.model,
|
||||
provider: this.model.provider,
|
||||
model: this.model.id,
|
||||
usage,
|
||||
stopReason,
|
||||
};
|
||||
} catch (error) {
|
||||
return {
|
||||
role: "assistant",
|
||||
model: this.model,
|
||||
provider: this.model.provider,
|
||||
model: this.model.id,
|
||||
usage: {
|
||||
input: 0,
|
||||
output: 0,
|
||||
|
|
@ -6,6 +6,7 @@ import type {
|
|||
LLM,
|
||||
LLMOptions,
|
||||
Message,
|
||||
Model,
|
||||
StopReason,
|
||||
TokenUsage,
|
||||
Tool,
|
||||
|
|
@ -19,9 +20,9 @@ export interface OpenAICompletionsLLMOptions extends LLMOptions {
|
|||
|
||||
export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
||||
private client: OpenAI;
|
||||
private model: string;
|
||||
private modelInfo: Model;
|
||||
|
||||
constructor(model: string, apiKey?: string, baseUrl?: string) {
|
||||
constructor(model: Model, apiKey?: string) {
|
||||
if (!apiKey) {
|
||||
if (!process.env.OPENAI_API_KEY) {
|
||||
throw new Error(
|
||||
|
|
@ -30,8 +31,12 @@ export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
|||
}
|
||||
apiKey = process.env.OPENAI_API_KEY;
|
||||
}
|
||||
this.client = new OpenAI({ apiKey, baseURL: baseUrl });
|
||||
this.model = model;
|
||||
this.client = new OpenAI({ apiKey, baseURL: model.baseUrl });
|
||||
this.modelInfo = model;
|
||||
}
|
||||
|
||||
getModel(): Model {
|
||||
return this.modelInfo;
|
||||
}
|
||||
|
||||
async complete(request: Context, options?: OpenAICompletionsLLMOptions): Promise<AssistantMessage> {
|
||||
|
|
@ -39,14 +44,14 @@ export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
|||
const messages = this.convertMessages(request.messages, request.systemPrompt);
|
||||
|
||||
const params: OpenAI.Chat.Completions.ChatCompletionCreateParamsStreaming = {
|
||||
model: this.model,
|
||||
model: this.modelInfo.id,
|
||||
messages,
|
||||
stream: true,
|
||||
stream_options: { include_usage: true },
|
||||
};
|
||||
|
||||
// Cerebras/xAI dont like the "store" field
|
||||
if (!this.client.baseURL?.includes("cerebras.ai") || this.client.baseURL?.includes("api.x.ai")) {
|
||||
if (!this.modelInfo.baseUrl?.includes("cerebras.ai") && !this.modelInfo.baseUrl?.includes("api.x.ai")) {
|
||||
(params as any).store = false;
|
||||
}
|
||||
|
||||
|
|
@ -66,7 +71,11 @@ export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
|||
params.tool_choice = options.toolChoice;
|
||||
}
|
||||
|
||||
if (options?.reasoningEffort && this.isReasoningModel() && !this.model.toLowerCase().includes("grok")) {
|
||||
if (
|
||||
options?.reasoningEffort &&
|
||||
this.modelInfo.reasoning &&
|
||||
!this.modelInfo.id.toLowerCase().includes("grok")
|
||||
) {
|
||||
params.reasoning_effort = options.reasoningEffort;
|
||||
}
|
||||
|
||||
|
|
@ -203,14 +212,16 @@ export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
|||
thinking: reasoningContent || undefined,
|
||||
thinkingSignature: reasoningField || undefined,
|
||||
toolCalls: toolCalls.length > 0 ? toolCalls : undefined,
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage,
|
||||
stopReason: this.mapStopReason(finishReason),
|
||||
};
|
||||
} catch (error) {
|
||||
return {
|
||||
role: "assistant",
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage: {
|
||||
input: 0,
|
||||
output: 0,
|
||||
|
|
@ -230,9 +241,9 @@ export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
|||
if (systemPrompt) {
|
||||
// Cerebras/xAi don't like the "developer" role
|
||||
const useDeveloperRole =
|
||||
this.isReasoningModel() &&
|
||||
!this.client.baseURL?.includes("cerebras.ai") &&
|
||||
!this.client.baseURL?.includes("api.x.ai");
|
||||
this.modelInfo.reasoning &&
|
||||
!this.modelInfo.baseUrl?.includes("cerebras.ai") &&
|
||||
!this.modelInfo.baseUrl?.includes("api.x.ai");
|
||||
const role = useDeveloperRole ? "developer" : "system";
|
||||
params.push({ role: role, content: systemPrompt });
|
||||
}
|
||||
|
|
@ -305,9 +316,4 @@ export class OpenAICompletionsLLM implements LLM<OpenAICompletionsLLMOptions> {
|
|||
return "stop";
|
||||
}
|
||||
}
|
||||
|
||||
private isReasoningModel(): boolean {
|
||||
// TODO base on models.dev
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -11,6 +11,7 @@ import type {
|
|||
LLM,
|
||||
LLMOptions,
|
||||
Message,
|
||||
Model,
|
||||
StopReason,
|
||||
TokenUsage,
|
||||
Tool,
|
||||
|
|
@ -24,9 +25,9 @@ export interface OpenAIResponsesLLMOptions extends LLMOptions {
|
|||
|
||||
export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
||||
private client: OpenAI;
|
||||
private model: string;
|
||||
private modelInfo: Model;
|
||||
|
||||
constructor(model: string, apiKey?: string, baseUrl?: string) {
|
||||
constructor(model: Model, apiKey?: string) {
|
||||
if (!apiKey) {
|
||||
if (!process.env.OPENAI_API_KEY) {
|
||||
throw new Error(
|
||||
|
|
@ -35,8 +36,12 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
}
|
||||
apiKey = process.env.OPENAI_API_KEY;
|
||||
}
|
||||
this.client = new OpenAI({ apiKey, baseURL: baseUrl });
|
||||
this.model = model;
|
||||
this.client = new OpenAI({ apiKey, baseURL: model.baseUrl });
|
||||
this.modelInfo = model;
|
||||
}
|
||||
|
||||
getModel(): Model {
|
||||
return this.modelInfo;
|
||||
}
|
||||
|
||||
async complete(request: Context, options?: OpenAIResponsesLLMOptions): Promise<AssistantMessage> {
|
||||
|
|
@ -44,7 +49,7 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
const input = this.convertToInput(request.messages, request.systemPrompt);
|
||||
|
||||
const params: ResponseCreateParamsStreaming = {
|
||||
model: this.model,
|
||||
model: this.modelInfo.id,
|
||||
input,
|
||||
stream: true,
|
||||
};
|
||||
|
|
@ -62,7 +67,7 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
}
|
||||
|
||||
// Add reasoning options for models that support it
|
||||
if (this.supportsReasoning() && (options?.reasoningEffort || options?.reasoningSummary)) {
|
||||
if (this.modelInfo?.reasoning && (options?.reasoningEffort || options?.reasoningSummary)) {
|
||||
params.reasoning = {
|
||||
effort: options?.reasoningEffort || "medium",
|
||||
summary: options?.reasoningSummary || "auto",
|
||||
|
|
@ -145,7 +150,8 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
else if (event.type === "error") {
|
||||
return {
|
||||
role: "assistant",
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage,
|
||||
stopReason: "error",
|
||||
error: `Code ${event.code}: ${event.message}` || "Unknown error",
|
||||
|
|
@ -159,14 +165,16 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
thinking: thinking || undefined,
|
||||
thinkingSignature: JSON.stringify(reasoningItems) || undefined,
|
||||
toolCalls: toolCalls.length > 0 ? toolCalls : undefined,
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage,
|
||||
stopReason,
|
||||
};
|
||||
} catch (error) {
|
||||
return {
|
||||
role: "assistant",
|
||||
model: this.model,
|
||||
provider: this.modelInfo.provider,
|
||||
model: this.modelInfo.id,
|
||||
usage: {
|
||||
input: 0,
|
||||
output: 0,
|
||||
|
|
@ -184,7 +192,7 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
|
||||
// Add system prompt if provided
|
||||
if (systemPrompt) {
|
||||
const role = this.supportsReasoning() ? "developer" : "system";
|
||||
const role = this.modelInfo?.reasoning ? "developer" : "system";
|
||||
input.push({
|
||||
role,
|
||||
content: systemPrompt,
|
||||
|
|
@ -260,14 +268,4 @@ export class OpenAIResponsesLLM implements LLM<OpenAIResponsesLLMOptions> {
|
|||
return "stop";
|
||||
}
|
||||
}
|
||||
|
||||
private supportsReasoning(): boolean {
|
||||
// TODO base on models.dev
|
||||
return (
|
||||
this.model.includes("o1") ||
|
||||
this.model.includes("o3") ||
|
||||
this.model.includes("gpt-5") ||
|
||||
this.model.includes("gpt-4o")
|
||||
);
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -8,29 +8,7 @@ export interface LLMOptions {
|
|||
|
||||
export interface LLM<T extends LLMOptions> {
|
||||
complete(request: Context, options?: T): Promise<AssistantMessage>;
|
||||
}
|
||||
|
||||
export interface ModelInfo {
|
||||
id: string;
|
||||
name: string;
|
||||
provider: string;
|
||||
capabilities: {
|
||||
reasoning: boolean;
|
||||
toolCall: boolean;
|
||||
vision: boolean;
|
||||
audio?: boolean;
|
||||
};
|
||||
cost: {
|
||||
input: number; // per million tokens
|
||||
output: number; // per million tokens
|
||||
cacheRead?: number;
|
||||
cacheWrite?: number;
|
||||
};
|
||||
limits: {
|
||||
context: number;
|
||||
output: number;
|
||||
};
|
||||
knowledge?: string;
|
||||
getModel(): Model;
|
||||
}
|
||||
|
||||
export interface UserMessage {
|
||||
|
|
@ -48,6 +26,7 @@ export interface AssistantMessage {
|
|||
name: string;
|
||||
arguments: Record<string, any>;
|
||||
}[];
|
||||
provider: string;
|
||||
model: string;
|
||||
usage: TokenUsage;
|
||||
|
||||
|
|
@ -112,6 +91,7 @@ export interface Model {
|
|||
id: string;
|
||||
name: string;
|
||||
provider: string;
|
||||
baseUrl?: string;
|
||||
reasoning: boolean;
|
||||
input: ("text" | "image")[];
|
||||
cost: {
|
||||
|
|
|
|||
|
|
@ -1,11 +1,11 @@
|
|||
import { describe, it, beforeAll, afterAll, expect } from "vitest";
|
||||
import { GoogleLLM } from "../src/providers/gemini.js";
|
||||
import { GoogleLLM } from "../src/providers/google.js";
|
||||
import { OpenAICompletionsLLM } from "../src/providers/openai-completions.js";
|
||||
import { OpenAIResponsesLLM } from "../src/providers/openai-responses.js";
|
||||
import { AnthropicLLM } from "../src/providers/anthropic.js";
|
||||
import type { LLM, LLMOptions, Context, Tool, AssistantMessage } from "../src/types.js";
|
||||
import type { LLM, LLMOptions, Context, Tool, AssistantMessage, Model } from "../src/types.js";
|
||||
import { spawn, ChildProcess, execSync } from "child_process";
|
||||
import { createLLM } from "../src/models.js";
|
||||
import { createLLM, getModel } from "../src/models.js";
|
||||
|
||||
// Calculator tool definition (same as examples)
|
||||
const calculatorTool: Tool = {
|
||||
|
|
@ -176,7 +176,7 @@ async function multiTurn<T extends LLMOptions>(llm: LLM<T>, thinkingOptions: T)
|
|||
const response = await llm.complete(context, thinkingOptions);
|
||||
context.messages.push(response);
|
||||
|
||||
if (response.content) {
|
||||
if (response.stopReason === "stop" && response.content) {
|
||||
finalResponse = response;
|
||||
break;
|
||||
}
|
||||
|
|
@ -217,7 +217,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: GoogleLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new GoogleLLM("gemini-2.5-flash", process.env.GEMINI_API_KEY!);
|
||||
llm = new GoogleLLM(getModel("google", "gemini-2.5-flash")!, process.env.GEMINI_API_KEY!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -245,7 +245,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: OpenAICompletionsLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new OpenAICompletionsLLM("gpt-4o-mini", process.env.OPENAI_API_KEY!);
|
||||
llm = new OpenAICompletionsLLM(getModel("openai", "gpt-4o-mini")!, process.env.OPENAI_API_KEY!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -265,7 +265,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: OpenAIResponsesLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new OpenAIResponsesLLM("gpt-5-mini", process.env.OPENAI_API_KEY!);
|
||||
llm = new OpenAIResponsesLLM(getModel("openai", "gpt-5-mini")!, process.env.OPENAI_API_KEY!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -293,7 +293,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: AnthropicLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new AnthropicLLM("claude-sonnet-4-0", process.env.ANTHROPIC_OAUTH_TOKEN!);
|
||||
llm = new AnthropicLLM(getModel("anthropic", "claude-sonnet-4-0")!, process.env.ANTHROPIC_OAUTH_TOKEN!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -321,7 +321,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: OpenAICompletionsLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new OpenAICompletionsLLM("grok-code-fast-1", process.env.XAI_API_KEY!, "https://api.x.ai/v1");
|
||||
llm = new OpenAICompletionsLLM(getModel("xai", "grok-code-fast-1")!, process.env.XAI_API_KEY!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -349,7 +349,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: OpenAICompletionsLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new OpenAICompletionsLLM("openai/gpt-oss-20b", process.env.GROQ_API_KEY!, "https://api.groq.com/openai/v1");
|
||||
llm = new OpenAICompletionsLLM(getModel("groq", "openai/gpt-oss-20b")!, process.env.GROQ_API_KEY!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -377,7 +377,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: OpenAICompletionsLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new OpenAICompletionsLLM("gpt-oss-120b", process.env.CEREBRAS_API_KEY!, "https://api.cerebras.ai/v1");
|
||||
llm = new OpenAICompletionsLLM(getModel("cerebras", "gpt-oss-120b")!, process.env.CEREBRAS_API_KEY!);
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -405,7 +405,7 @@ describe("AI Providers E2E Tests", () => {
|
|||
let llm: OpenAICompletionsLLM;
|
||||
|
||||
beforeAll(() => {
|
||||
llm = new OpenAICompletionsLLM("z-ai/glm-4.5", process.env.OPENROUTER_API_KEY!, "https://openrouter.ai/api/v1");
|
||||
llm = new OpenAICompletionsLLM(getModel("openrouter", "z-ai/glm-4.5")!, process.env.OPENROUTER_API_KEY!);;
|
||||
});
|
||||
|
||||
it("should complete basic text generation", async () => {
|
||||
|
|
@ -479,7 +479,23 @@ describe("AI Providers E2E Tests", () => {
|
|||
setTimeout(checkServer, 1000); // Initial delay
|
||||
});
|
||||
|
||||
llm = new OpenAICompletionsLLM("gpt-oss:20b", "dummy", "http://localhost:11434/v1");
|
||||
const model: Model = {
|
||||
id: "gpt-oss:20b",
|
||||
provider: "ollama",
|
||||
baseUrl: "http://localhost:11434/v1",
|
||||
reasoning: true,
|
||||
input: ["text"],
|
||||
contextWindow: 128000,
|
||||
maxTokens: 16000,
|
||||
cost: {
|
||||
input: 0,
|
||||
output: 0,
|
||||
cacheRead: 0,
|
||||
cacheWrite: 0,
|
||||
},
|
||||
name: "Ollama GPT-OSS 20B"
|
||||
}
|
||||
llm = new OpenAICompletionsLLM(model, "dummy");
|
||||
}, 30000); // 30 second timeout for setup
|
||||
|
||||
afterAll(() => {
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue