atlas
trn1…
Try

3 of 1225 instances · cheapest is trn1.2xlarge at $981/mo

Instance Workload vCPUs Memory Local disk Network On-demand/mo ↑ Per vCPU
trn1.2xlargeGPU & accelerators 8 32 GiB 475 GB NVMe 12.5 Gbps $981 $123
trn1.32xlargeGPU & accelerators 128 512 GiB 7.6 TB NVMe 800 Gbps $15,695 $123
trn1n.32xlargeGPU & accelerators 128 512 GiB 7.6 TB NVMe 1600 Gbps $18,089 $141