32K

Out

Chat

disable

DBRX Instruct

DBRX Instruct API: Databricks' state-of-the-art open large language model with advanced features and superior performance on industry benchmarks.

Free $1 Tokens for New Members

Text to Speech

Javascript

Python

                                        const { OpenAI } = require('openai');

const api = new OpenAI({
  baseURL: 'https://api.ai.cc/v1',
  apiKey: '',
});

const main = async () => {
  const result = await api.chat.completions.create({
    model: 'databricks/dbrx-instruct',
    messages: [
      {
        role: 'system',
        content: 'You are an AI assistant who knows everything.',
      },
      {
        role: 'user',
        content: 'Tell me, why is the sky blue?'
      }
    ],
  });

  const message = result.choices[0].message.content;
  console.log(`Assistant: ${message}`);
};

main();

                                        import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.ai.cc/v1",
    api_key="",    
)

response = client.chat.completions.create(
    model="databricks/dbrx-instruct",
    messages=[
        {
            "role": "system",
            "content": "You are an AI assistant who knows everything.",
        },
        {
            "role": "user",
            "content": "Tell me, why is the sky blue?"
        },
    ],
)

message = response.choices[0].message.content

print(f"Assistant: {message}")

Docs

One API 300+ AI Models

Save 20% on Costs & $1 Free Tokens

Get API Key Explore Models

DBRX Instruct

Product Detail

💻 Introducing DBRX Instruct: A New Era of Open LLMs

DBRX Instruct, developed by Databricks, is a groundbreaking large language model (LLM) designed to set new benchmarks in performance and efficiency. Released in 2023, this version 1.0 Instruct model leverages a sophisticated architecture to deliver superior capabilities across a wide spectrum of natural language processing tasks.

● Model Name: DBRX
● Developer/Creator: Databricks
● Release Date: 2023
● Version: 1.0 Instruct
● Model Type: Large Language Model (LLM)

🔥 Key Architectural & Performance Highlights

At its core, DBRX Instruct utilizes a fine-grained Mixture-of-Experts (MoE) architecture. This innovative design incorporates 132 billion total parameters, with a dynamic activation of 36 billion parameters for any given input, ensuring optimal efficiency and performance.

✨ Core Features:

✓ Advanced MoE System: Features 16 experts capable of selecting 4, offering 65x more possible expert combinations than other prominent open MoE models.
✓ Extensive Training Data: Pre-trained on an impressive 12 trillion tokens of meticulously curated text and code data.
✓ Benchmark Dominance: Demonstrates exceptional performance across critical benchmarks including general knowledge, commonsense reasoning, programming, and mathematical reasoning.
✓ Outperforms Peers: Consistently surpasses leading open models such as Mixtral Instruct and Code Llama (70B) in various evaluations.

📜 Intended Use Cases & Multilingual Support

DBRX Instruct is engineered as a general-purpose LLM, making it incredibly versatile for a multitude of natural language processing (NLP) applications.

💬 Ideal For:

✍ Text Generation: Crafting coherent and contextually relevant text.
❓ Question Answering: Providing accurate and insightful responses to queries.
💻 Code Generation: Generating high-quality code snippets and solving programming challenges.
🔢 Mathematical Reasoning: Excelling in tasks requiring complex mathematical understanding.

Furthermore, DBRX Instruct stands out as a multilingual model, capable of processing and generating content across a broad spectrum of languages, enhancing its global applicability.

🔗 Deep Dive: Technical Specifications & Performance

Architecture

DBRX Instruct is built upon a transformer-based, decoder-only LLM architecture, trained using the next-token prediction objective. Its fine-grained MoE setup involves 16 distinct experts, dynamically selecting 4 for each input query to optimize processing.

Training Data Quality

The model's robust capabilities stem from its pre-training on 12 trillion tokens of meticulously curated text and code data. With a maximum context length of 32,000 tokens, this dataset is estimated to be at least twice the quality of data used for the MPT family of models, ensuring a rich understanding and generation capacity.

📈 Performance Metrics vs. Leading Models:

DBRX Instruct consistently demonstrates superior performance against other leading open models on standard benchmarks:

MMLU: 73.7% (DBRX Instruct) vs. 71.4% (Mixtral Instruct)
HellaSwag 10-shot: 89.0% (DBRX Instruct) vs. 87.6% (Mixtral Instruct)
WinoGrande: 81.8% (DBRX Instruct) vs. 81.1% (Mixtral Instruct)
Databricks Gauntlet: 66.8% (DBRX Instruct) vs. 60.7% (Mixtral Instruct)
HumanEval: 70.1% (DBRX Instruct) vs. 54.8% (Mixtral Instruct)
GSM8k: 66.9% (DBRX Instruct) vs. 61.1% (Mixtral Instruct)

💡 Getting Started with DBRX Instruct

API Access

DBRX Instruct is designed for easy integration via API. An example API snippet (e.g., `open-ai.chat-completion` with `databricks/dbrx-instruct`) demonstrates its straightforward usability for developers.

📝 Licensing Information

DBRX Instruct is available for use under the Databricks Open Model License, promoting broad access and innovation.

💬 Frequently Asked Questions (FAQ)

Q: What is the DBRX Instruct model?

A: DBRX Instruct is a powerful, open-source large language model (LLM) developed by Databricks, known for its fine-grained Mixture-of-Experts (MoE) architecture and strong performance across various NLP tasks.
Q: How does DBRX Instruct differ from other LLMs?

A: It uses a unique MoE architecture with 16 experts (selecting 4 per input), offering significantly more expert combinations and outperforming leading open models like Mixtral Instruct and Code Llama (70B) on key benchmarks.
Q: What are the primary applications of DBRX Instruct?

A: It's a general-purpose LLM ideal for text generation, question answering, code generation, and tasks requiring strong programming and mathematical reasoning capabilities.
Q: Is DBRX Instruct multilingual?

A: Yes, DBRX Instruct supports a wide range of languages, making it suitable for global applications.
Q: Under what license is DBRX Instruct available?

A: DBRX Instruct is released under the Databricks Open Model License.

AI Playground

Test all API models in the sandbox environment before you integrate. We provide more than 300 models to integrate into your app.

Try For Free

One API
300+ AI Models

Save 20% on Costs

Free $1 Tokens for New Members