Grok 4 Fast (As of 11/8/2025) is now the best cheap reasoning model with its latency, speed, intelligence and token per second this would be my go to model for most tasks if it were available. Please consider adding it.