Verified90% confidenceFactTime unknown
The M4 chip's Neural Engine can perform 38 trillion operations per second (38 TOPS)
11
Sources
90%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedApple began integrating a neural engine starting with the A11 chip, and through multiple generations of evolution in the M series, performance has improved from an initial hundreds of billions of operations per second to tens of trillions.73% similarUnverifiedThe CS4 launched by Cerebras is claimed to be the fastest AI inference hardware currently available, ultimately capable of running models with up to 100 trillion parameters at a speed of approximately 1,000 tokens per second.71% similarUnverifiedThe Neural Engine integrated in the M2 Ultra contains 32 dedicated cores with a theoretical compute performance of up to 31.6 TOPS.69% similarVerifiedApple's Neural Engine first appeared with the A11 Bionic chip in 2017, with the A11 capable of executing 600 billion operations per second, while the A18 Pro chip has reached 35 trillion operations per second.69% similarUnverifiedNVIDIA engineers, by writing a brand-new dedicated fused decoding kernel upstream for VLLM, achieved a throughput of 423 tokens per second per user on the 280-billion-parameter Kimi K3 model, a 1.68x speedup on equivalent hardware.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/29659API
curl https://kongchang.com/api/v1/knowledge/claims/29659MCP
get_claim(id=29659)