Baseten vs Replicate
As of , Baseten lists the lower on-demand rate on 2 of the 3 GPUs both offer on-demand; Replicate wins on 1. On the priciest shared card both list on-demand, NVIDIA H100 80GB (form factor not stated): Baseten $6.50 vs Replicate $5.49 per GPU-hour on-demand — Replicate is $1.01 (16%) cheaper.
On-demand, per GPU-hour
| GPU |  Baseten |  Replicate | Difference | Cheaper |
|---|---|---|---|---|
| NVIDIA H100 80GB (form factor not stated) median $3.75 · 28 providers · lowest Gcore $1.72 | $6.50 captured Deploy ↗ | $5.49 captured Deploy ↗ | $1.01 (16%) | Replicate |
| NVIDIA A100 80GB median $1.99 · 63 providers · lowest Yotta Labs $0.92 | $4.00 captured Deploy ↗ | $5.04 captured Deploy ↗ | $1.04 (21%) | Baseten |
| NVIDIA T4 median $0.55 · 19 providers · lowest immers.cloud $0.24 | $0.63 captured Deploy ↗ | $0.81 captured Deploy ↗ | $0.18 (22%) | Baseten |
How this is compared
Cloud GPU rental prices of Baseten and Replicate, compared per GPU-hour on the GPUs both list. Where a provider has several tiers or regions, its cheapest on-demand rate is used and labelled. Community capacity (RunPod Community Cloud) and marketplace floors are shown only where a provider has nothing else, and never decide which is cheaper.
GPU by GPU
- NVIDIA H100 80GB (form factor not stated): Baseten $6.50 vs Replicate $5.49 per GPU-hour on-demand — Replicate is $1.01 (16%) cheaper.
- NVIDIA A100 80GB: Baseten $4.00 vs Replicate $5.04 per GPU-hour on-demand — Baseten is $1.04 (21%) cheaper.
- NVIDIA T4: Baseten $0.63 vs Replicate $0.81 per GPU-hour on-demand — Baseten is $0.18 (22%) cheaper.
Not compared: GPUs only one of the two lists, reserved/committed tiers, and anything either provider prices only on request. See each provider's page for its full list.
Data updated: