BREAKING
Grok 4.5 Undercuts Rivals on Cost
0$
Grok 4.5 per task
0%
cheaper than Fable 5
Intelligence Index Scores
Fable 560
GPT-5.6 Sol59
Opus 4.856
Grok 4.554
Cost Per Task Compared
Fable 52.75
Opus 4.81.8
Grok 4.50.31
Value Pick With Trade-Offs
StrengthsValue
60% fewer output tokens than Opus 4.8
Approaches Fable 5 on coding agents
Cheap for heavy agentic use
CautionsWatch
Lags Fable 5 on hardest coding
Higher hallucination risk
Leaner token budget
Pricing Over Peak Benchmarks
AI NEWS BLITZ
xAI's Grok 4.5 is reframing the frontier-model race around cost, not just raw benchmarks.