models.ts

import { ILLM, ModelProvider } from "core";
import { ModelProviderTags } from "../../../components/modelSelection/utils";
import { InputDescriptor } from "./providers";

// A dimension is like parameter count - 7b, 13b, 34b, etc.
// You would set options to the field that should be changed for that option in the params field of ModelPackage
export interface PackageDimension {
  name: string;
  description: string;
  options: { [key: string]: { [key: string]: any } };
}
export interface ModelPackage {
  title: string;
  icon?: string;
  collectInputFor?: InputDescriptor[];
  description: string;
  refUrl?: string;
  tags?: ModelProviderTags[];
  params: {
    model: ILLM["model"];
    templateMessages?: ILLM["templateMessages"];
    contextLength: ILLM["contextLength"];
    stopTokens?: string[];
    promptTemplates?: ILLM["promptTemplates"];
    replace?: [string, string][];
    [key: string]: any;
  };
  dimensions?: PackageDimension[];
  providerOptions?: ModelProvider[];
  isOpenSource: boolean;
}

export const models: { [key: string]: ModelPackage } = {
  granite34bCodeInstruct: {
    title: "Granite 34B Code Instruct",
    description: "A highly capable code instruct model with 34 billion parameters by IBM",
    params: {
      title: "Granite 34B Code Instruct",
      model: "granite-34b-code-instruct",
      contextLength: 8192,
      apiKey: "<API_KEY>",
    },
    icon: "ibm.png",
    providerOptions: ["ibm"],
    isOpenSource: true,
  },
  granite8bCodeInstruct: {
    title: "Granite 8B Code Instruct",
    description: "A highly capable code instruct model with 8 billion parameters by IBM",
    params: {
      title: "Granite 8B Code Instruct",
      model: "granite-8b-code-instruct",
      contextLength: 4096,
      apiKey: "<API_KEY>",
    },
    icon: "ibm.png",
    providerOptions: ["ibm"],
    isOpenSource: true,
  },
  granite3bCodeInstruct: {
    title: "Granite 3B Code Instruct",
    description: "A highly capable code instruct model with 3 billion parameters by IBM",
    params: {
      title: "Granite 3B Code Instruct",
      model: "granite-3b-code-instruct",
      contextLength: 2048,
      apiKey: "<API_KEY>",
    },
    icon: "ibm.png",
    providerOptions: ["ibm"],
    isOpenSource: true,
  },
  llama31Chat: {
    title: "Llama3.1 Chat",
    description: "A model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.1-8b",
      model: "llama3.1-8b",
      contextLength: 8192,
    },
    icon: "meta.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "8b": {
            model: "llama3.1-8b",
            title: "Llama3.1-8b",
          },
          "70b": {
            model: "llama3.1-70b",
            title: "Llama3.1-70b",
          },
          "405b": {
            model: "llama3.1-405b",
            title: "Llama3.1-405b",
          },
        },
      },
    ],
    providerOptions: [
      "ollama",
      "lmstudio",
      "together",
      "llama.cpp",
      "replicate",
      "sambanova",
      "cerebras",
      "nebius",
    ],
    isOpenSource: true,
  },
  llama32Chat: {
    title: "Llama3.2 Chat",
    description:
      "The latest model from Meta, fine-tuned for chat. Llama3.1 recommended - chat stayed the same",
    refUrl: "",
    params: {
      title: "Llama3.2-11b",
      model: "llama3.2-11b",
      contextLength: 8192,
    },
    icon: "meta.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "1b": {
            model: "llama3.2-1b",
            title: "Llama3.2-1b",
          },
          "3b": {
            model: "llama3.2-3b",
            title: "Llama3.2-3b",
          },
          "11b": {
            model: "llama3.2-11b",
            title: "Llama3.2-11b",
          },
          "90b": {
            model: "llama3.2-90b",
            title: "Llama3.2-90b",
          },
        },
      },
    ],
    providerOptions: ["ollama", "groq", "llama.cpp"],
    isOpenSource: true,
  },
  deepseek: {
    title: "DeepSeek Coder",
    description:
      "A model pre-trained on 2 trillion tokens including 80+ programming languages and a repo-level corpus.",
    params: {
      title: "DeepSeek-7b",
      model: "deepseek-7b",
      contextLength: 4096,
    },
    icon: "deepseek.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "1b": {
            model: "deepseek-1b",
            title: "DeepSeek-1b",
          },
          "7b": {
            model: "deepseek-7b",
            title: "DeepSeek-7b",
          },
          "33b": {
            model: "deepseek-33b",
            title: "DeepSeek-33b",
          },
        },
      },
    ],
    providerOptions: ["ollama", "lmstudio", "llama.cpp"],
    isOpenSource: true,
  },
  deepseekChatApi: {
    title: "DeepSeek Chat",
    description: "DeepSeek's best model for general chat use cases.",
    params: {
      title: "DeepSeek Chat",
      model: "deepseek-chat",
      contextLength: 128_000,
    },
    icon: "deepseek.png",
    providerOptions: ["deepseek"],
    isOpenSource: false,
  },
  deepseekCoderApi: {
    title: "DeepSeek Coder",
    description:
      "A model pre-trained on 2 trillion tokens including 80+ programming languages and a repo-level corpus.",
    params: {
      title: "DeepSeek Coder",
      model: "deepseek-coder",
      contextLength: 128_000,
    },
    icon: "deepseek.png",
    providerOptions: ["deepseek"],
    isOpenSource: false,
  },
  deepseekCoder2Lite: {
    title: "DeepSeek Coder 2 Lite",
    description:
      "DeepSeek-Coder-V2-Lite-Instruct is an open-source code language model that supports 338 programming languages and offers a context length of up to 128,000 tokens.",
    params: {
      title: "DeepSeek-2-lite",
      model: "deepseek-2-lite",
      contextLength: 128_000,
    },
    icon: "deepseek.png",
    providerOptions: ["nebius"],
    isOpenSource: true,
  },
  moonshotChat: {
    title: "Moonshot Chat",
    description: "Moonshot AI provides high-performance large language models",
    refUrl: "https://platform.moonshot.cn/",
    params: {
      title: "Moonshot-v1-8k",
      model: "moonshot-v1-8k",
      contextLength: 8192,
    },
    icon: "moonshot.png",
    dimensions: [
      {
        name: "Context Window",
        description: "The size of the model's context window",
        options: {
          "8K": {
            model: "moonshot-v1-8k",
            title: "Moonshot-v1-8k",
            contextLength: 8192,
          },
          "32K": {
            model: "moonshot-v1-32k",
            title: "Moonshot-v1-32k",
            contextLength: 32768,
          },
          "128K": {
            model: "moonshot-v1-128k",
            title: "Moonshot-v1-128k",
            contextLength: 131072,
          },
        },
      },
    ],
    providerOptions: ["moonshot"],
    isOpenSource: false,
  },
  mistralOs: {
    title: "Mistral",
    description:
      "A series of open-weight models created by Mistral AI, highly competent for code generation and other tasks",
    params: {
      title: "Mistral",
      model: "mistral-7b",
      contextLength: 4096,
    },
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "7b": {
            model: "mistral-7b",
            title: "Mistral-7b",
          },
          "8x7b (MoE)": {
            model: "mistral-8x7b",
            title: "Mixtral",
          },
          "8x22b (MoE)": {
            model: "mistral-8x22b",
            title: "Mixtral",
          },
        },
      },
    ],
    icon: "mistral.png",
    providerOptions: [
      "ollama",
      "lmstudio",
      "together",
      "llama.cpp",
      "replicate",
      "nebius",
    ],
    isOpenSource: true,
  },
  codeLlamaInstruct: {
    title: "CodeLlama Instruct",
    description:
      "A model from Meta, fine-tuned for code generation and conversation",
    refUrl: "",
    params: {
      title: "CodeLlama-7b",
      model: "codellama-7b",
      contextLength: 4096,
    },
    icon: "meta.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "7b": {
            model: "codellama-7b",
            title: "CodeLlama-7b",
          },
          "13b": {
            model: "codellama-13b",
            title: "CodeLlama-13b",
          },
          "34b": {
            model: "codellama-34b",
            title: "CodeLlama-34b",
          },
          "70b": {
            model: "codellama-70b",
            title: "Codellama-70b",
          },
        },
      },
    ],
    providerOptions: [
      "ollama",
      "lmstudio",
      "together",
      "llama.cpp",
      "replicate",
    ],
    isOpenSource: true,
  },
  llama3170bTrial: {
    title: "Codellama 70b (Free Trial)",
    description:
      "The best code model from Meta, fine-tuned for code generation and conversation",
    refUrl: "",
    params: {
      title: "CodeLlama-70b",
      model: "codellama-70b",
      contextLength: 4096,
    },
    icon: "meta.png",
    providerOptions: ["free-trial"],
    isOpenSource: false,
  },
  llama31405bTrial: {
    title: "Llama3.1 405b (Free Trial)",
    description: "The latest Llama model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.1-405b",
      model: "llama3.1-405b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["free-trial"],
    isOpenSource: false,
  },
  mixtralTrial: {
    title: "Mixtral (Free Trial)",
    description:
      "Mixtral 8x7b is a mixture of experts model created by Mistral AI",
    refUrl: "",
    params: {
      title: "Mixtral",
      model: "mistral-8x7b",
      contextLength: 4096,
    },
    icon: "mistral.png",
    providerOptions: ["free-trial", "groq"],
    isOpenSource: false,
  },
  llama38bChat: {
    title: "Llama3 8b",
    description: "The latest Llama model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3-8b",
      model: "llama3-8b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["groq"],
    isOpenSource: false,
  },
  llama370bChat: {
    title: "Llama3 70b Chat",
    description: "The latest Llama model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3-70b",
      model: "llama3-70b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["groq", "askSage"],
    isOpenSource: false,
  },
  llama318bChat: {
    title: "Llama3.1 8b Chat",
    description: "A model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.1-8b",
      model: "llama3.1-8b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["groq"],
    isOpenSource: false,
  },
  llama3170bChat: {
    title: "Llama3.1 70b Chat",
    description: "A model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.1-70b",
      model: "llama3.1-70b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["groq"],
    isOpenSource: false,
  },
  llama31405bChat: {
    title: "Llama3.1 405b Chat",
    description: "A model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.1-405b",
      model: "llama3.1-405b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["groq"],
    isOpenSource: false,
  },
  llama3170bNemotron: {
    title: "Llama 3.1 Nemotron 70B",
    description:
      "Llama-3.1-Nemotron-70B-Instruct is a large language model customized by NVIDIA to improve the helpfulness of LLM generated responses to user queries.",
    params: {
      title: "Llama 3.1 Nemotron 70B",
      model: "llama3.1-70b-nemotron",
      contextLength: 128_000,
    },
    icon: "meta.png",
    isOpenSource: true,
  },
  llama321bChat: {
    title: "Llama3.2 1b Chat",
    description:
      "The latest super-lightweight model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.2-1b",
      model: "llama3.2-1b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["ollama", "groq", "llama.cpp", "sambanova"],
    isOpenSource: false,
  },
  llama323bChat: {
    title: "Llama3.2 3b Chat",
    description: "The latest lightweight model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3.2-3b",
      model: "llama3.2-3b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["ollama", "groq", "llama.cpp", "sambanova", "together"],
    isOpenSource: false,
  },
  llama3211bChat: {
    title: "Llama3.2 11b Chat",
    description: "The latest lightweight multi-modal model from Meta",
    refUrl: "",
    params: {
      title: "Llama3.2-11b",
      model: "llama3.2-11b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["ollama", "groq", "llama.cpp", "together"],
    isOpenSource: false,
  },
  llama3290bChat: {
    title: "Llama3.2 90b Chat",
    description: "The latest lightweight multi-modal model from Meta",
    refUrl: "",
    params: {
      title: "Llama3.2-90b",
      model: "llama3.2-90b",
      contextLength: 8192,
    },
    icon: "meta.png",
    providerOptions: ["ollama", "groq", "llama.cpp"],
    isOpenSource: false,
  },
  llama3Chat: {
    title: "Llama3 Chat",
    description: "A model from Meta, fine-tuned for chat",
    refUrl: "",
    params: {
      title: "Llama3-8b",
      model: "llama3-8b",
      contextLength: 8192,
    },
    icon: "meta.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "8b": {
            model: "llama3-8b",
            title: "Llama3-8b",
          },
          "70b": {
            model: "llama3-70b",
            title: "Llama3-70b",
          },
        },
      },
    ],
    providerOptions: [
      "ollama",
      "lmstudio",
      "together",
      "llama.cpp",
      "replicate",
      "nebius",
    ],
    isOpenSource: true,
  },
  graniteCodeOpenSource: {
    title: "Granite Code",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "granite-code",
      contextLength: 20_000,
      title: "Granite Code",
      systemMessage: `You are Granite Chat, an AI language model developed by IBM. You are a cautious assistant. You carefully follow instructions. You are helpful and harmless and you follow ethical guidelines and promote positive behavior. You always respond to greetings (for example, hi, hello, g'day, morning, afternoon, evening, night, what's up, nice to meet you, sup, etc) with "Hello! I am Granite Chat, created by IBM. How can I help you today?". Please do not say anything else and do not start a conversation.`,
    },
    providerOptions: ["ollama", "lmstudio", "llama.cpp", "replicate"],
    icon: "ibm.png",
    isOpenSource: true,
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "3b": {
            model: "granite-code-3b",
            title: "Granite Code 3B",
          },
          "8b": {
            model: "granite-code-8b",
            title: "Granite Code 8B",
          },
          "20b": {
            model: "granite-code-20b",
            title: "Granite Code 20B",
          },
          "34b": {
            model: "granite-code-34b",
            title: "Granite Code 34B",
          },
        },
      },
    ],
  },
  wizardCoder: {
    title: "WizardCoder",
    description:
      "A CodeLlama-based code generation model from WizardLM, focused on Python",
    refUrl: "",
    params: {
      title: "WizardCoder-7b",
      model: "wizardcoder-7b",
      contextLength: 4096,
    },
    icon: "wizardlm.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "7b": {
            model: "wizardcoder-7b",
            title: "WizardCoder-7b",
          },
          "13b": {
            model: "wizardcoder-13b",
            title: "WizardCoder-13b",
          },
          "34b": {
            model: "wizardcoder-34b",
            title: "WizardCoder-34b",
          },
        },
      },
    ],
    providerOptions: ["ollama", "lmstudio", "llama.cpp", "replicate"],
    isOpenSource: true,
  },
  phindCodeLlama: {
    title: "Phind CodeLlama (34b)",
    description: "A finetune of CodeLlama by Phind",
    icon: "meta.png",
    params: {
      title: "Phind CodeLlama",
      model: "phind-codellama-34b",
      contextLength: 4096,
    },
    providerOptions: [
      "ollama",
      "lmstudio",
      "llama.cpp",
      "replicate",
      "free-trial",
    ],
    isOpenSource: true,
  },
  codestral: {
    title: "Codestral",
    description:
      "Codestral is an advanced generative model created by Mistral AI, tailored for coding tasks like fill-in-the-middle and code completion.",
    params: {
      title: "Codestral",
      model: "codestral-latest",
      contextLength: 32000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral"],
    isOpenSource: true,
  },
  codestralMamba: {
    title: "Codestral Mamba",
    description: "A Mamba 2 language model specialized in code generation.",
    params: {
      title: "Codestral Mamba",
      model: "codestral-mamba-latest",
      contextLength: 256_000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral"],
    isOpenSource: true,
  },
  mistral7b: {
    title: "Mistral 7B",
    description:
      "The first dense model released by Mistral AI, perfect for experimentation, customization, and quick iteration. At the time of the release, it matched the capabilities of models up to 30B parameters.",
    params: {
      title: "Mistral 7B",
      model: "open-mistral-7b",
      contextLength: 32000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral"],
    isOpenSource: false,
  },
  mistral8x7b: {
    title: "Mixtral 8x7B",
    description:
      "A sparse mixture of experts model. As such, it leverages up to 45B parameters but only uses about 12B during inference, leading to better inference throughput at the cost of more vRAM.",
    params: {
      title: "Mixtral 8x7B",
      model: "open-mixtral-8x7b",
      contextLength: 32000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral"],
    isOpenSource: false,
  },
  mistral8x22b: {
    title: "Mistral 8x22B",
    description:
      "A bigger sparse mixture of experts model. As such, it leverages up to 141B parameters but only uses about 39B during inference, leading to better inference throughput at the cost of more vRAM.",
    params: {
      title: "Mistral 8x22B",
      model: "open-mixtral-8x22b",
      contextLength: 64000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral"],
    isOpenSource: false,
  },
  mistralSmall: {
    title: "Mistral Small",
    description:
      "Suitable for simple tasks that one can do in bulk (Classification, Customer Support, or Text Generation)",
    params: {
      title: "Mistral Small",
      model: "mistral-small-latest",
      contextLength: 32000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral"],
    isOpenSource: false,
  },
  mistralLarge: {
    title: "Mistral Large",
    description:
      "Mistral's flagship model that's ideal for complex tasks that require large reasoning capabilities or are highly specialized (Synthetic Text Generation, Code Generation, RAG, or Agents).",
    params: {
      title: "Mistral Large",
      model: "mistral-large-latest",
      contextLength: 32000,
    },
    icon: "mistral.png",
    providerOptions: ["mistral", "askSage"],
    isOpenSource: false,
  },
  mistralNemo: {
    title: "Mistral Nemo",
    description:
      "Mistral Nemo Instruct is a large language model developed by Mistral AI and NVIDIA.",
    params: {
      title: "Mistral Nemo",
      model: "mistral-nemo",
      contextLength: 128_000,
    },
    icon: "mistral.png",
    isOpenSource: true,
  },
  geminiPro: {
    title: "Gemini Pro",
    description: "A highly capable model created by Google DeepMind",
    params: {
      title: "Gemini Pro",
      model: "gemini-pro",
      contextLength: 32_000,
      apiKey: "<API_KEY>",
    },
    icon: "gemini.png",
    providerOptions: ["gemini"],
    isOpenSource: false,
  },
  gemini15Pro: {
    title: "Gemini 1.5 Pro",
    description: "A newer Gemini model with 1M token context length",
    params: {
      title: "Gemini 1.5 Pro",
      model: "gemini-1.5-pro-latest",
      contextLength: 2_000_000,
      apiKey: "<API_KEY>",
    },
    icon: "gemini.png",
    providerOptions: ["gemini", "free-trial", "askSage"],
    isOpenSource: false,
  },
  gemini15Flash: {
    title: "Gemini 1.5 Flash",
    description:
      "Fast and versatile multimodal model for scaling across diverse tasks",
    params: {
      title: "Gemini 1.5 Flash",
      model: "gemini-1.5-flash-latest",
      contextLength: 1_000_000,
      apiKey: "<API_KEY>",
    },
    icon: "gemini.png",
    providerOptions: ["gemini"],
    isOpenSource: false,
  },
  commandR: {
    title: "Command R",
    description:
      "Command R is a scalable generative model targeting RAG and Tool Use to enable production-scale AI for enterprise.",
    params: {
      model: "command-r",
      contextLength: 128_000,
      title: "Command R",
      apiKey: "",
    },
    providerOptions: ["cohere"],
    icon: "cohere.png",
    isOpenSource: false,
  },
  commandRPlus: {
    title: "Command R+",
    description:
      "Command R+ is a state-of-the-art RAG-optimized model designed to tackle enterprise-grade workloads",
    params: {
      model: "command-r-plus",
      contextLength: 128_000,
      title: "Command R+",
      apiKey: "",
    },
    providerOptions: ["cohere"],
    icon: "cohere.png",
    isOpenSource: false,
  },
  gpt4turbo: {
    title: "GPT-4 Turbo",
    description:
      "A faster and more capable version of GPT-4 with longer context length and image support",
    params: {
      model: "gpt-4-turbo",
      contextLength: 128_000,
      title: "GPT-4 Turbo",
    },
    providerOptions: ["openai"],
    icon: "openai.png",
    isOpenSource: false,
  },
  gpt4o: {
    title: "GPT-4o",
    description:
      "An even faster version of GPT-4 with stronger multi-modal capabilities.",
    params: {
      model: "gpt-4o",
      contextLength: 128_000,
      title: "GPT-4o",
      systemMessage:
        "You are an expert software developer. You give helpful and concise responses.",
    },
    providerOptions: ["openai", "free-trial", "askSage"],
    icon: "openai.png",
    isOpenSource: false,
  },
  gpt4omini: {
    title: "GPT-4o Mini",
    description:
      "A model at less than half the price of gpt-3.5-turbo, but near gpt-4 in capabilities.",
    params: {
      model: "gpt-4o-mini",
      contextLength: 128_000,
      title: "GPT-4o mini",
      systemMessage:
        "You are an expert software developer. You give helpful and concise responses.",
    },
    providerOptions: ["openai", "askSage"],
    icon: "openai.png",
    isOpenSource: false,
  },
  gpt35turbo: {
    title: "GPT-3.5-Turbo",
    description:
      "A faster, cheaper OpenAI model with slightly lower capabilities",
    params: {
      model: "gpt-3.5-turbo",
      contextLength: 8096,
      title: "GPT-3.5-Turbo",
    },
    providerOptions: ["openai", "free-trial", "askSage"],
    icon: "openai.png",
    isOpenSource: false,
  },
  claude35Sonnet: {
    title: "Claude 3.5 Sonnet",
    description:
      "Anthropic's most intelligent model, but much less expensive than Claude 3 Opus",
    params: {
      model: "claude-3-5-sonnet-latest",
      contextLength: 200_000,
      title: "Claude 3.5 Sonnet",
      apiKey: "",
    },
    providerOptions: ["anthropic", "free-trial", "askSage"],
    icon: "anthropic.png",
    isOpenSource: false,
  },
  claude3Opus: {
    title: "Claude 3 Opus",
    description:
      "The most capable model in the Claude 3 series, beating GPT-4 on many benchmarks",
    params: {
      model: "claude-3-opus-20240229",
      contextLength: 200_000,
      title: "Claude 3 Opus",
      apiKey: "",
    },
    providerOptions: ["anthropic", "askSage"],
    icon: "anthropic.png",
    isOpenSource: false,
  },
  claude3Sonnet: {
    title: "Claude 3 Sonnet",
    description:
      "The second most capable model in the Claude 3 series: ideal balance of intelligence and speed",
    params: {
      model: "claude-3-sonnet-20240229",
      contextLength: 200_000,
      title: "Claude 3 Sonnet",
      apiKey: "",
    },
    providerOptions: ["anthropic", "free-trial", "askSage"],
    icon: "anthropic.png",
    isOpenSource: false,
  },
  claude3Haiku: {
    title: "Claude 3 Haiku",
    description:
      "The third most capable model in the Claude 3 series: fastest and most compact model for near-instant responsiveness",
    params: {
      model: "claude-3-haiku-20240307",
      contextLength: 200_000,
      title: "Claude 3 Haiku",
      apiKey: "",
    },
    providerOptions: ["anthropic", "free-trial"],
    icon: "anthropic.png",
    isOpenSource: false,
  },
  graniteChat: {
    title: "Granite Chat 13b",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-13b-chat-v2",
      contextLength: 8_000,
      title: "Granite Chat 13b",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  graniteCode3b: {
    title: "Granite Code 3b",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-3b-code-instruct",
      contextLength: 2_000,
      title: "Granite Code 3b",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  graniteCode8b: {
    title: "Granite Code 8b",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-8b-code-instruct",
      contextLength: 4_000,
      title: "Granite Code 8b",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  graniteCode20b: {
    title: "Granite Code 20b",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-20b-code-instruct",
      contextLength: 8_000,
      title: "Granite Code 20b",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  graniteCode34b: {
    title: "Granite Code 34b",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-34b-code-instruct",
      contextLength: 8_000,
      title: "Granite Code 34b",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  granite3Instruct8b: {
    title: "Granite 3.0 8b Instruct",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-3-8b-instruct",
      contextLength: 8_000,
      title: "Granite 3.0 8b Instruct",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  granite3Instruct2b: {
    title: "Granite 3.0 2b Instruct",
    description:
      "The Granite model series is a family of IBM-trained, dense decoder-only models, which are particularly well-suited for generative tasks.",
    params: {
      model: "ibm/granite-3-2b-instruct",
      contextLength: 2_000,
      title: "Granite 3.0 2b Instruct",
    },
    providerOptions: ["watsonx"],
    icon: "WatsonX.png",
    isOpenSource: false,
  },
  MistralLarge: {
    title: "Mistral Large",
    description:
      "Mistral Large, the most advanced Large Language Model (LLM) developed by Mistral Al, is an exceptionally powerful model.",
    params: {
      model: "mistralai/mistral-large",
      contextLength: 20_000,
      title: "Mistral Large",
    },
    providerOptions: ["watsonx"],
    icon: "mistral.png",
    isOpenSource: false,
  },
  MetaLlama3: {
    title: "Llama 3.1",
    description:
      "Llama 3 is an auto-regressive language model that uses an optimized transformer architecture.",
    params: {
      title: "Llama 3.1",
      model: "meta-llama/llama-3-1-8b-instruct",
      contextLength: 20_000,
    },
    icon: "meta.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "8b": {
            model: "meta-llama/llama-3-1-8b-instruct",
            title: "Llama 3.1 8b",
          },
          "70b": {
            model: "meta-llama/llama-3-1-70b-instruct",
            title: "Llama 3.1 70b",
          },
        },
      },
    ],
    providerOptions: ["watsonx"],
    isOpenSource: false,
  },
  VertexGemini15Pro: {
    title: "Gemini 1.5 Pro",
    description: "A newer Gemini model with 1M token context length",
    params: {
      title: "Gemini 1.5 Pro",
      model: "gemini-1.5-pro-002",
      contextLength: 2_097_152,
    },
    icon: "gemini.png",
    providerOptions: ["vertexai"],
    isOpenSource: false,
  },
  VertexGemini15Flash: {
    title: "Gemini 1.5 Flash",
    description:
      "Fast and versatile multimodal model for scaling across diverse tasks",
    params: {
      title: "Gemini 1.5 Flash",
      model: "gemini-1.5-flash-002",
      contextLength: 1_048_576,
    },
    icon: "gemini.png",
    providerOptions: ["vertexai"],
    isOpenSource: false,
  },
  vertexMistralLarge: {
    title: "Mistral Large",
    description:
      "Mistral's flagship model that's ideal for complex tasks that require large reasoning capabilities or are highly specialized (Synthetic Text Generation, Code Generation, RAG, or Agents).",
    params: {
      title: "Mistral Large",
      model: "mistral-large",
      contextLength: 32000,
    },
    icon: "mistral.png",
    providerOptions: ["vertexai"],
    isOpenSource: false,
  },
  gpt4gov: {
    title: "GPT-4 gov",
    description:
      "U.S. Government. Most capable model today - which is similar to GPT-4o but approved for use by the U.S. Government.",
    params: {
      model: "gpt4-gov",
      contextLength: 128_000,
      title: "GPT-4 gov",
      systemMessage:
        "You are an expert software developer. You give helpful and concise responses.", // Need to set this on the Ask Sage side or just configure it in here to be discussed
    },
    providerOptions: ["askSage"],
    icon: "openai.png",
    isOpenSource: false,
  },
  gpt4ogov: {
    title: "GPT-4o gov",
    description:
      "U.S. Government. Most capable model today - which is similar to GPT-4o but approved for use by the U.S. Government.",
    params: {
      model: "gpt-4o-gov",
      contextLength: 128_000,
      title: "GPT-4o",
      systemMessage:
        "You are an expert software developer. You give helpful and concise responses.", // Need to set this on the Ask Sage side or just configure it in here to be discussed
    },
    providerOptions: ["askSage"],
    icon: "openai.png",
    isOpenSource: false,
  },
  Qwen2Coder: {
    title: "Qwen 2.5 Coder 7b",
    description:
      "Qwen 2.5 is an auto-regressive language model that uses an optimized transformer architecture.",
    params: {
      title: "Qwen 2.5 Coder 7b",
      model: "qwen-coder2.5-7b",
      contextLength: 32_000,
    },
    icon: "qwen.png",
    dimensions: [
      {
        name: "Parameter Count",
        description: "The number of parameters in the model",
        options: {
          "7b": {
            model: "qwen-coder2.5-7b",
            title: "Qwen 2.5 Coder 7b",
          },
        },
      },
    ],
    providerOptions: ["nebius"],
    isOpenSource: true,
  },
  grokBeta: {
    title: "Grok Beta",
    description: "Generative artificial intelligence chatbot developed by xAI.",
    refUrl: "",
    params: {
      title: "Grok Beta",
      model: "grok-beta",
      contextLength: 128_000,
    },
    icon: "xAI.png",
    providerOptions: ["xAI"],
    isOpenSource: false,
  },
  gemma2_2b: {
    title: "Gemma 2 2b IT",
    description:
      "Gemma 2 IT instruction-tuned language model developed by Google, designed for tasks like question answering, summarization, and reasoning.",
    params: {
      title: "Gemma 2 2b IT",
      model: "gemma2-2b-it",
      contextLength: 4000,
    },
    isOpenSource: true,
  },
  gemma2_9b: {
    title: "Gemma 2 9b IT",
    description:
      "Gemma 2 IT instruction-tuned language model developed by Google, designed for tasks like question answering, summarization, and reasoning.",
    params: {
      title: "Gemma 2 9b IT",
      model: "gemma2-9b-it",
      contextLength: 8000,
    },
    isOpenSource: true,
  },
  phi3mini: {
    title: "Microsoft Phi 3 mini",
    description:
      "Phi 3 Mini is a 3.8-billion-parameter language model developed by Microsoft.",
    params: {
      title: "Microsoft Phi 3 mini",
      model: "phi-3-mini",
      contextLength: 4000,
    },
    isOpenSource: true,
  },
  phi3medium: {
    title: "Microsoft Phi 3 medium",
    description:
      "Phi 3 Medium is a 14-billion-parameter language model developed by Microsoft.",
    params: {
      title: "Microsoft Phi 3 medium",
      model: "phi-3-medium",
      contextLength: 128_000,
    },
    isOpenSource: true,
  },
  olmo7b: {
    title: "OLMo 7b",
    description:
      "OLMo 7B Instruct HF is a 7-billion-parameter language model by the Allen Institute for AI, fine-tuned for question answering tasks.",
    params: {
      title: "OLMo 7b",
      model: "olmo-7b",
      contextLength: 2000,
    },
    isOpenSource: true,
  },
  AUTODETECT: {
    title: "Autodetect",
    description:
      "Automatically populate the model list by calling the /models endpoint of the server",
    params: {
      model: "AUTODETECT",
    } as any,
    providerOptions: [],
    isOpenSource: false,
  },
};