19 August 2026

AI inference speed claims jump sharply across multiple vendors

Alibaba's Qwen model reportedly generated 70 tokens per second on Apple's M5 Max chip, showing faster processing on consumer devices.
Cerebras announced a CS-4 system claiming 1000 tokens per second for large models, a significant jump in datacenter performance.
Faster inference, meaning quicker AI response times, now affects user experience, operating costs, and how governments view AI competitiveness.

How it was covered

DFlash 2 claims Qwen3.8-27B at 70 tokens per second on M5 Max while Cerebras announced CS-4 with claims of 1000 tokens per second for 10T models. Inference speed is becoming product UX, economics, and national competitiveness policy.