Well, it’s been a hot minute (okay, exactly a year): xAI just dropped the weights for Grok 2! 🚀 Same architecture as Grok 1 – MoE with eight experts on deck. Just a reminder, its metrics are hanging out around the same level as GPT-4o and Claude 3.5 Sonnet. This checkpoint? A hefty 500 GB! You’ll need a solid setup with 8 GPUs, each rocking over 40 GB of memory. But hey, good news! The license allows for free non-commercial AND commercial use up to $1 million in annual revenue. 💰 Check it out on Hugging Face! 🤖✨
Well, it’s been a hot minute (okay, exactly a year): xAI just dropped the weights for Grok 2! 🚀
Same architecture as Grok 1 – MoE with eight experts on deck. Just a reminder, its metrics are hanging out around the same level as GPT-4o and Claude 3.5 Sonnet.
This checkpoint? A hefty 500 GB! You’ll need a solid setup with 8 GPUs, each rocking over 40 GB of memory. But hey, good news! The license allows for free non-commercial AND commercial use up to $1 million in annual revenue. 💰
Check it out on Hugging Face! 🤖✨
