← All benchmarks
Headline Methodology Results Raw

5 MAY 2026 · VPS50 · cloud · 16 vCPU AMD EPYC · 62 GB RAM · no GPU · gemma/phi/qwen2.5/qwen3 · param-sweep

vps50-cpu-matrix-1 — gemma/phi/qwen2.5/qwen3 on vps50

TL;DR

0 request records in the retained JSONL. No timed request records retained. Only setup/resource or microbenchmark records are available. Incomplete evidence: this record cannot establish chat latency or generation performance.

Methodology

Request counts were checked against the retained JSONL on 21 October 2026, excluding telemetry and skipped records. Timing aggregates remain those published in the source. Count audit and original summary counts.

See Public methodology and measurement limits for the full procedure. Recorded source revision (where available): —.

Results

Cell tok/s mean tok/s p50 tok/s p95 duration p50 calls
phi-4————0
gemma-4-26b-a4b————0
qwen3-30b-a3b————0
qwen2.5-72b————0
gemma-4-26b-a4b-it-q4km-cpu-ctx32k————0

tokens per second — mean · p50 · p95

No tokens-per-second data captured.

Raw data

Every run gets its JSONL, log, summary, and metadata published. Clone the archive; re-run it; tell us where we got it wrong.

Cite

Margetic, S. et al. (2026). benchmarks.weeyuga.com/benchmarks/91751afd.html
Public benchmarks of the Weeyuga cluster. Run id: 91751afd-068a-477b-8f40-6e1963f803f1. SHA —.

Related runs