GPU Cloud Picks by Use Case
Opinionated recommendations across 30 providers, organized by what you're actually trying to do. Each page splits the recommendation three ways — performance, ops, cost — because the cheapest provider is rarely the easiest to operate, and the fastest GPU is rarely the cheapest per token.
Live pricing + 30-day reliability scores update on every page load. Curation refreshed manually — last updated May 2026.
Training
Training picks split three ways. The cost-optimal provider isn't always the performance-optimal one — multi-node training is interconnect-bound, and free egress matters more than the headline $/hr when you're moving terabytes of checkpoints. Here's how three lenses (performance, ops, cost) actually shake out, with live pricing.
View 3 picks →
Inference
Inference economics differ from training: latency, cold-start, billing granularity, and per-token cost all matter more than peak FLOPS. The right GPU for an 8B model is rarely the right GPU for a 405B model. Here's how the three axes split.
View 3 picks →
Fine-tuning
Fine-tuning sits between training and inference: short bursts of GPU work interleaved with eval, dataset shuffling, and human review. Billing granularity (per-second vs per-hour), persistent storage, and dataset egress hit harder than the headline $/hr. Three picks across the axes.
View 3 picks →
Budget
Cheap doesn't mean cheap. A $1.80/hr H100 at 70% reliability costs more in re-runs than a $2.39/hr H100 at 95%. The picks below show the absolute floor, the most reliable budget option, and the ones to actively avoid despite their brand recognition.
View 3 picks →
Consumer / Local LLM
If you'd otherwise drop $2,000 on a 4090 to run Llama or Qwen at home, renting a consumer card by the hour is often the smarter move: no upfront capex, no power bill, and you can size up to a 5090's 32 GB or down to a 3090's 24 GB per job. The three axes below cover the fastest consumer card, the best day-to-day workhorse, and the absolute cheapest 24 GB you can rent. These are GeForce-class cards — great for single-GPU inference and light fine-tuning, not for multi-GPU training.
View 3 picks →