Grok 4.7 Token Efficiency Falls 30-80% Short of Claims, Real-World Costs 2x Grok 4.6

flowersslop · x · 2026-09-22

theo's hands-on review: Grok 4.5 was a great value default model, but Grok 4.6 got slower and pricier with far higher token usage for marginal gains. Grok 4.7 is harder to forgive — despite claims of better token efficiency, it's actually 30-80% worse, scores below 4.6 on benchmarks, and real-world costs exceed 2x Grok 4.6, above Astra. Reposter flowersslop notes the irony: Elon said he no longer cares about benchmarks, yet is now posting niche ones again.

Original post →

More from Models

Models channel →