BREAKING
Grok 4.6 Launches
0
Intelligence Index
0
USD/M
Input tokens
0
USD/M
Output tokens
Five Points Over Grok 4.5
Grok 4.5 High
56
Grok 4.6 High
61
Designed for Long Agent Runs
1
Research
↓
2
Build
↓
3
Self-test
↓
4
Refine
Strong Gains, Open Questions
Reported strengths
Early use
●
Coding and planning
●
Competitive token pricing
Remaining limits
To verify
●
Production reliability
●
Verbose reasoning and tool use
Production Tests Come Next
AI NEWS BLITZ
SpaceXAI is releasing a new model built for sustained agent work.