EN

Enflame Yunsui T20 vs Enflame L600

Side-by-side specifications, performance, and rental pricing for datacenter AI workloads.

Yunsui T20: GCU / DTU 2.0, 2021 L600: GCU-CARA 4, 2025
Memory
32 vs 144 GB
112 GB more for the L600
TDP
300 W vs 700 W
Yunsui T20 draws 400 W less

Specifications

EF
Yunsui T20
EF
L600
Architecture
GCU / DTU 2.0GCU-CARA 4
Launch Year
20212025
Form Factor
PCIeOAM
Memory
32 GB144 GB
Memory Bandwidth
——
TDP
300 W700 W
Process Node
GlobalFoundries 12nm—

Performance (TFLOPS)

EF
Yunsui T20
EF
L600
FP64
No verified data available No verified data available
FP32
34 TFLOPS No verified data available
TF32
134 TFLOPS No verified data available
BF16
134 TFLOPS No verified data available
FP16
134 TFLOPS No verified data available
FP8
No verified data available No verified data available
FP6
No verified data available No verified data available
FP4
No verified data available No verified data available
INT8
269 TOPS No verified data available

FLOPS by Precision

What actually differs

The L600 is the newer part: GCU-CARA 4, launched in 2025, against the Yunsui T20's GCU / DTU 2.0 from 2021. Newer architectures typically add lower-precision formats and better throughput per watt, so check the precision rows your workload actually uses.

Memory is 32 GB against 144 GB. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Which has more memory, the Yunsui T20 or the L600?

The L600 carries 144 GB of memory versus 32 GB for the Yunsui T20.

How much power do the Yunsui T20 and the L600 draw?

The Yunsui T20 is rated at 300 W TDP and the L600 at 700 W. On FP32 throughput per watt, the Yunsui T20 is the more efficient part.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

© 2026 Flopper.io - Compare the hardware powering AI