Elon Musk’s XAI Open Releases Grok-1 Architecture

Elon Musk’s X.AI Corp., doing business as xAI, released the base model weights and network architecture of its artificial intelligence or AI chatbot Grok-1.

The open sourcing of Grok-1 would allow researchers and developers to build on the large language model.

Grok-1 is a 314 billion parameter Mixture-of-Experts model trained from scratch by xAI, the AI startup founded by Musk in March 2023.

Musk in November 2023 had unveiled xAI’s first AI chatbot technology named Grok, that competes with OpenAI’s highly popular Generative AI ChatGPT.

xAI then noted that Grok is an AI modeled after the Hitchhiker’s Guide to the Galaxy, and is intended to answer almost anything and, far harder, even suggest what questions to ask.

Grok has real-time knowledge of the world through the X platform, formerly Twitter, and will also answer spicy questions that are rejected by most other AI systems.

The company now said it is releasing the weights and the architecture under the Apache 2.0 license.

XAI said in its announcement, “This is the raw base model checkpoint from the Grok-1 pre-training phase, which concluded in October 2023. This means that the model is not fine-tuned for any specific application, such as dialogue.”

The company noted that it was trained from scratch by xAI using a custom training stack on top of JAX and Rust in October 2023.

The interested parties can get started with using the model by following the instructions at github.com/xai-org/grok.

Musk last week had stated that xAI, which competes with AI developers OpenAI, Google, Meta, and others, would open source its AI chatbot.

Musk, who was the founding member of OpenAI, recently filed a lawsuit against OpenAI and its CEO Sam Altman alleging that the firm is developing AI for profitability instead of its foundational mission of utilizing it for the benefit of humankind.

OpenAI responded to Musk’s lawsuit by publishing the latter’s emails acknowledging that it needs to raise a ton of money to fulfil its AI missions.

OpenAI reportedly said recently that it would seek to dismiss the suit.

In recent developments, Google in February had issued an apology for its AI chatbot Gemini for generating images of historical figures with inaccurate racial and ethnic depictions. The tech major attributed the errors to its attempt to create diverse results.

Business News