Load
The checkpoint alone requires at least 13 GPUs at this reserve, before runtime allocations.
PLANNING ENGINE / PUBLIC PREVIEW
Test an exact checkpoint against an explicit hardware topology with sourced bytes and inspectable integer math.
A baseline pass is a candidate—not a claim about runtime support or performance.
Deterministic registry engine
Test an exact, pinned artifact against advertised accelerator memory—then keep every unsupported conclusion visibly unknown.
02 · Fit decision
Kimi K3 · 896 / top-16 experts · 1M context · 8 × NVIDIA H200 SXM · 13% reserve
The checkpoint alone requires at least 13 GPUs at this reserve, before runtime allocations.
A baseline pass does not prove loader, kernel, quantization, sharding, or expert-parallel support.
No measured workload profile is attached, so throughput, latency, KV demand, and skew are not projected.
No dated provider, region, or utilization record is selected, so the engine emits no invented cost.
03 · Explainable topology
RequestedThe topology you asked the engine to test against the checkpoint baseline.
usable = floor(advertised bytes × (10,000 − reserve bps) / 10,000); minimum GPUs = ceil(checkpoint tensor bytes / usable bytes). A failure is conclusive for this no-offload baseline; a pass is only a candidate for runtime validation.
Artifact evidence: Kimi K3 pinned tensor manifest ↗
Exact revision and Safetensors manifest—not parameter-count arithmetic.
Integer basis-point math preserves a visible memory assumption.
Runtime, KV, scale, and price stay unknown without attached evidence.
The same registry and calculation ship through the open CLI.