gpudiff the public record of change in GPU cloud pricing
Follow the diffs: RSS · Free API · Sponsor this site

What's the cheapest GPU that fits my model?

Single-GPU fit: parameters × bytes-per-weight × 1.2 overhead, against nameplate VRAM and live prices. Rough by design — KV cache scales with context, and multi-GPU sharding changes everything. A screening tool, not a capacity plan.

Needs:
GPUVRAM GBOn-demand $/hrWhereSpot from