bench.de

bench.de is a command for coding agents from druide67/asiai. It costs 38 tokens per session (1,820 once invoked), scanned B, original, Apache-2.0.

A command-line benchmark for comparing language-model engines on Apple Silicon Macs. It measures output speed, first-response delay, and energy use with standard prompts.

In plain words
What is it for?
Use it to benchmark coding, tool-use, reasoning, or long-output prompts; compare models; fill different context sizes; view benchmark history; export JSON; and create or share result cards.
Why use it?
It replaces informal impressions with comparable measurements across engines or models.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/druide67/asiai/bench.de
Clone the repo
git clone --depth 1 https://github.com/druide67/asiai

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for bench.de

README.md
[![agentmods](https://agentmods.dev/badge/commands/druide67/asiai/bench.de.svg)](https://agentmods.dev/commands/druide67/asiai/bench.de)
Your own site
<a href="https://agentmods.dev/commands/druide67/asiai/bench.de"><img src="https://agentmods.dev/badge/commands/druide67/asiai/bench.de.svg" alt="Measured on agentmods" height="20"></a>
Per session 38 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,820 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00038 $0.01820
Opus 5 $0.00019 $0.00910
Sonnet 5 $0.00008 $0.00364
Haiku 4.5 $0.00004 $0.00182

Measured 3d ago against content hash e0ce7b40af22, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

bench.de scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

| `--power` | Kreuzvalidierung der Leistung mit sudo powermetrics (IOReport immer aktiv) |
docs/commands/bench.de.md · 170 lines

How it starts

The opening of the file, as written. The whole thing — 170 lines — stays where its author put it; the contents beside it link to each section on GitHub.

asiai bench

Engine-übergreifender Benchmark mit standardisierten Prompts.

Verwendung

asiai bench [options]

Optionen

Option Beschreibung
-m, --model MODEL Zu benchmarkendes Modell (Standard: Auto-Erkennung)
-e, --engines LIST Engines filtern (z.B. ollama,lmstudio,mlxlm)
-p, --prompts LIST Prompt-Typen: code, tool_call, reasoning, long_gen
-r, --runs N Durchläufe pro Prompt (Standard: 3, für Median + Stddev)
--power Kreuzvalidierung der Leistung mit sudo powermetrics (IOReport immer aktiv)
--context-size SIZE Kontext-Füll-Prompt: 4k, 16k, 32k, 64k
--export FILE Ergebnisse als JSON-Datei exportieren
-H, --history PERIOD Vergangene Benchmarks anzeigen (z.B. 7d, 24h)
-Q, --quick Schnellbenchmark: 1 Prompt (code), 1 Durchlauf (~15 Sekunden)
--compare MODEL [MODEL...] Modellübergreifender Vergleich (2-8 Modelle, gegenseitig exklusiv mit -m)
--card Teilbare Benchmark-Karte generieren (SVG lokal, PNG mit --share)
--share Ergebnisse in die Community-Benchmark-Datenbank übermitteln

Beispiel

asiai bench -m qwen3.5 --runs 3 --power
  Mac Mini M4 Pro — Apple M4 Pro  RAM: 64.0 GB (42% used)  Pressure: normal

Benchmark: qwen3.5

  Engine       tok/s (±stddev)    Tokens   Duration     TTFT       VRAM    Thermal
  ────────── ───────────────── ───────── ────────── ──────── ────────── ──────────
  lmstudio    72.6 ± 0.0 (stable)   435    6.20s    0.28s        —    nominal
  ollama      30.4 ± 0.1 (stable)   448   15.28s    0.25s   26.0 GB   nominal

  Winner: lmstudio (2.4x faster)
  Power: lmstudio 13.2W (5.52 tok/s/W) — ollama 16.0W (1.89 tok/s/W)

Prompts

Vier standardisierte Prompts testen verschiedene Generierungsmuster:

Name Tokens Testet
code 512 Strukturierte Code-Generierung (BST in Python)
tool_call 256 JSON-Funktionsaufrufe / Instruktionsbefolgung
reasoning 384 Mehrstufiges Mathematikproblem
long_gen 1024 Dauerhafter Durchsatz (Bash-Skript)

Read the full file on GitHub · 170 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 170 lines · 38 tokens per session scan B e0ce7b40af22

Subscribe to this mod's changes

bench.de is a command published in the GitHub repository druide67/asiai (11 stars, last pushed 4d ago), licensed Apache-2.0. It adds 38 tokens to every session and 1,820 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.