xAI, Elon Musk’s AI company, has unveiled the Grok-1.5 model, boasting the ability to support messages up to 128,000 tokens in length. This new model has demonstrated impressive performance on benchmark tests, achieving 50.6% on MATH (4-shot) and an outstanding 90% on GSM8K (8-shot). It has proven its superiority over Mistral Large model from Mistral and even the recently distributed Grok-1 model, yet still falls short compared to Claude 3 Opus and GPT-4 in several test sets.
xAI has also announced that the Grok-1.5 model will soon be available for use via ๐ and is currently hiring Infra maintenance staff.
TLDR: xAI introduces the powerful Grok-1.5 model with enhanced text support capabilities and impressive benchmark test results, gearing up for future applications and expansion.
Leave a Comment