WSE-3
WSE-3 is a Cerebras wafer-scale AI processor based on the Wafer-scale SRAM mesh architecture, first released in 2024.
Built from BenchGecko data as of April 14, 2026 · updates when the data changes
WSE-3 is a Cerebras wafer-scale AI processor based on the Wafer-scale SRAM mesh architecture, first released in 2024.
Basic
The Cerebras Wafer-Scale Engine 3 is a wafer-scale AI processor made by Cerebras. Peak throughput is 125,000 TFLOPS at FP16/BF16. It carries 44 GB of SRAM on-wafer with 21,000 TB/s of memory bandwidth. Rated power (TDP) is 23,000 W. It is built on N5 at TSMC.
Deep
In the BenchGecko dataset it sits in the specialized tier: wafer-scale, LPU and other custom designs. Disclosed buyers include G42. Specs come from manufacturer datasheets; the spec card on this page shows the dataset values and their as-of date.
Expert
That is 5.43 FP16 TFLOPS per watt. Memory bandwidth per unit of compute is 168 GB/s per FP16 TFLOP: a higher ratio helps memory-bound inference, while peak TFLOPS matter more for compute-bound training. Fabrication: Taiwan. Real throughput depends on the model, precision, batch size and software stack, so compare tokens per second per dollar for your own workload.
Depending on why you're here
- ·WSE-3: 125,000 TFLOPS FP16/BF16
- ·Process: N5 at TSMC
- ·Tier: specialized
- ·44 GB of SRAM on-wafer per chip
- ·Cloud availability in the dataset: Cerebras
- ·Compare it with other chips on /hardware/wse-3
- ·Disclosed buyers: G42
- ·Cerebras · fabbed at TSMC
- ·Supply tightness is tracked on /hardware
- ·WSE-3 is a chip built to run AI, made by Cerebras
- ·It has 44 GB of fast memory so large models fit
- ·Companies buy or rent these to train and serve AI models
Frequently Asked Questions
Read the primary sources
- WSE-3 product pagewww.cerebras.ai