BREAKING
xAI Releases Grok 4.5
0K
context tokens
0$
input / M tokens
0$
output / M tokens
Grok 4.5 vs Claude Opus
Grok 4.5MoE
Comparable to Opus
Far fewer tokens used
Cached at $0.50
Claude Opus
Higher token usage
Benchmark reference
0days
to first operation
0k
initial H100 GPUs
0M+
GPU target
From beta to broad use
1SpaceX/Tesla beta
2Public release
3Cursor & Databricks
Power may be the bottleneck
AI NEWS BLITZ
xAI has officially launched Grok 4.5 for coding, agentic tasks and knowledge work.