Evidence · benchmarks · download
Proof you can open — no NDA required for the headlines.
Every material number on this site points here: public packs, measured tables, and demo binaries.
Benchmarks
Results from public LuxiDemo evidence on H100 hardware (2026-07-11). Open any pack to inspect methods and raw traces.
7B-class TRADE — energy & throughput
Hardware: 1× NVIDIA H100 NVL · 28-layer stack · public pack.
| Sequence length | Tokens / sec | Joules / token | Median power (W) | Notes |
|---|---|---|---|---|
| 5 | ~44 | 3.560 ± 0.005 | ~156 | Short-burst reference |
| 32 | ~221 | — | — | Throughput sweep |
| 64 | ~359 | — | — | Throughput sweep |
| 128 | ~403 | 0.630 ± 0.002 | ~254 | Primary result |
| 256 | ~464 | ~0.604 | ~280 | Longer context |
| 512 | ~493 | — | — | Throughput sweep |
Primary multi-run at sequence length 128. Source: evidence/h100-7b-class-TRADE/
Head-to-head with a Flash baseline
12 layers · sequence 1024 · h=768 · ~10 s each side.
| Metric | TRADE 12L | PyTorch + Flash 12L | Ratio |
|---|---|---|---|
| Median power (W) | 177.2 | 176.4 | ~1.0× |
| ms / stack forward | 74.32 | 3.90 | TRADE 19.1× slower |
| Prefill tok/s | 13778 | 262859 | PT 19.1× higher |
| Joules / token | 0.0129 | 0.0007 | TRADE 19.2× higher |
h100-stack12-H2H — we publish where a standard baseline wins on this shape.
Long-context memory
| Sequence length | Waller state (MB) | Dense scores (MB) | Reduction |
|---|---|---|---|
| 1,024 | 0.52 | 8.4 | 16× |
| 4,096 | 2.1 | 134 | 64× |
| 8,192 | 4.2 | 537 | 128× |
| 32,768 | 16.8 | 4,295 | 256× |
Memory scales near-linearly vs dense O(N²). Pack →
Evidence packs
Open on GitHub. Read methods. Check traces.
Lead packh100-7b-class-TRADE
Full 28-layer 7B-class TRADE stack. Multi-run thr + energy (~0.63 J/tok @ seq=128, ~403 tok/s).
Energyh100-stack12-TRADE-cuda
Device-resident 12-layer stack energy (GPT-2 width).
Honestyh100-stack12-H2H
TRADE vs PyTorch Flash — we publish the loss on this shape.
AUDITh100-WNSM-free-ride
Null-space free-ride under GPU load.
Memoryh100-LONGCTX-scaling
O(N) vs O(N²) memory ladder + CUDA 32k.
Microh100-BASELINE-vs-geo
Single-layer baseline wedges — not a 7B thr claim.
Serveh100-serve-sustain-2026-07-11
Continuous-batch serve sustain with power traces.
Download
Public demo builds for the expression/receipt path. Demo only — not production inference. Energy claims live in the packs above, not only inside these binaries.
Open latest release →LuxiEdge Lite / Demo
macOS ARM, Linux x86 (CPU/GPU), Linux ARM, Windows — from the public release channel.
Quick start
chmod +x luxiedge-*-macos-arm64
./luxiedge-*-macos-arm64 --port 9090
curl -X POST http://localhost:9090/evaluate \
-H "Content-Type: application/json" \
-d '{"expr":"sin(x)","values":[0.5,1.0],"precision":"f32"}'