Unverified50% confidenceFactExact time
The Grok model series has iterated from Grok-1 to Grok-2 and then to Grok-3, with Grok-3 claiming to have been trained using more than 100,000 H100 GPUs when it was released in early 2025.
1
Sources
50%
Confidence
Long-term
Relevance
9/4/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedGrok 4.5 was trained on tens of thousands of NVIDIA GB300 GPUs, supports an ultra-long context of up to 500,000 tokens, and natively supports visual input.77% similarUnverifiedGrok 4.5 was trained on tens of thousands of NVIDIA GB300 GPUs.72% similarUnverifiedGrok is based on a proprietary Grok-1 model with a parameter scale of 314 billion, using a Mixture of Experts (MoE) architecture.69% similarUnverifiedxAI officially stated that Grok 4.5 is the most powerful model to date and the first version in the Grok series positioned as a general-purpose capability model that goes beyond software engineering.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/852965API
curl https://kongchang.com/api/v1/knowledge/claims/852965MCP
get_claim(id=852965)