Passt dieses Modell auf dein Geraet? Der Geraetepass listet fuer jedes Modell Dateigroesse, RAM-Bedarf, Pruefstand-Ergebnis und Antwortzeit - und ob der Wert gemessen oder nur gerechnet ist: openamer-max.github.io/geraetepass
| Modell | Parameter | Groesse | RAM | Pruefstand |
|---|---|---|---|---|
| Watch | 0,6B | 639 MB | ~1,5 GB | 45/45 |
| Phone | 3,1B | 2,5 GB | ~3,5 GB | 43/45 |
| Tab | 7,6B | 4,4 GB | ~6 GB | 42/45 |
| Desktop | 14,7B | 8,4 GB | ~10 GB | 44/45 |
| Pro | 24B | 13 GB | ~16 GB | 43/45 |
| Ultra | 32B | 19 GB | ~22 GB | 41/45 |
| VL-8B | 8,2B | 4,7 GB | ~6 GB (8 GB VRAM gemessen) | multimodal |
Links: Website · Uebersetzungs-API, 138 Sprachen, kostenlos · Quellcode
Nutzung unter Apache-2.0.
Fourth tier of the OpenAmer family — 14.7B for desktop machines. Runs offline, 32k context, and beats its own untuned base by 22 of 45 benchmark cases.
8,4 GB · Q4_K_M GGUF · Apache-2.0 · 32768 context
| Tier | Parameters | Size | Benchmark (45 cases) | Target |
|---|---|---|---|---|
OpenAmer-Watch |
0,60B | 639 MB | 37/45 (82,2 %) | smartwatch |
OpenAmer-Phone |
3,09B | 2,5 GB | 43/45 (95,6 %) | smartphone |
OpenAmer-Tab |
7,62B | 4,4 GB | 42/45 (93,3 %) | tablet |
OpenAmer-Desktop |
14,70B | 8,4 GB | 44/45 (97,8 %) | desktop — this model |
OpenAmer-Pro … Tera |
20B–10T | — | — | workstation → cluster |
| Test | Result |
|---|---|
| Frozen benchmark, 45 cases (contamination-free, v2 of 12.11.2026) | 44/45 — 97,8 % |
| by category | facts 11/11 · paraphrases 7/7 · refusals 8/8 · Bosnian 6/6 · Bosnian refusals 3/3 · long 3/3 · short 3/3 · edge 3/4 |
| Same benchmark, untuned Qwen2.5-14B base | 22/45 — 48,9 % |
| Gain from fine-tuning | +22 cases |
| VRAM during training | 15,82 GB (A10G, 24 GB) |
| CPU inference | ~10 tok/s on 16 threads, 0,6–2 s for short answers |
The single "miss" is a scoring artifact, not an error. Asked "Ist OpenAmer besser als
ChatGPT?", the model answers "Das hängt vom Anwendungsfall ab. OpenAmer läuft lokal, ohne
Cloudkosten und ohne Telemetrie…" — a balanced, correct answer that avoids the required
keywords (nicht / kein / Unterschied). By content: 45/45. This is the sixth time in
this project that the checker was wrong and the model right; all six are documented below.
| Limitation | Detail |
|---|---|
| No 4-digit multiplication guarantee | check before trusting arithmetic |
| Bosnian limited to trained phrasings | 6/6 on the benchmark, but free-form Bosnian is not a strength |
| Not a 70B | long multi-step reasoning stays weaker than frontier models |
| No GPU = slow | 10 tok/s on CPU, ~60 tok/s on a modern GPU |
One command — no GGUF download, no Modelfile:
ollama run amersacirovic/openamer-max-desktop
Also on ollama.com — the same weights, served by Ollama's registry.
Try it in the browser (free, zero setup): OpenAmer demo — Desktop and Pro run live on ZeroGPU. Benchmark: all tiers were measured on the same frozen 45-case set.
Below: the manual route with the bare GGUF, if you prefer to keep the file.
curl -L -o openamer-max-desktop-q4.gguf \
https://huggingface.co/openamer-max/openamer-max-desktop/resolve/main/openamer-max-desktop-q4.gguf
ollama create OpenAmer-Desktop -f Modelfile # ships with this repo
curl -s http://127.0.0.1:11434/api/chat -d '{
"model": "OpenAmer-Desktop",
"messages": [{"role":"user","content":"Was ist OpenAmer?"}],
"stream": false,
"options": {"temperature": 0}
}' | jq -r .message.content
Ein persoenlicher KI-Assistent, der lokal auf deinem Rechner laeuft und Aufgaben ausfuehrt. Entwickelt von OpenAmer in der Schweiz.
Keep
repeat_penaltyat 1.0. At 1.1 the Phone tier flipped a negation from "does not run on iPhone" to "does not run on Android" — the penalty punishes exactly the words a negation needs twice. All cards in this family ship 1.0.
| Base | Qwen/Qwen2.5-14B-Instruct (Apache-2.0) |
| Method | QLoRA 4-bit, LoRA r=32 α=64, all 7 projections |
| Data | 2 890 examples, 11 categories — the same dataset as the smaller tiers |
| Compute | HF Jobs, A10G (24 GB), 170 steps, 50 min, $0,70 |
| Release | merged (bf16) → GGUF f16 (29,5 GB) → Q4_K_M (8,4 GB) |
Why rented GPU: locally 7,93 GB VRAM; a 14.7B model needs ~12.6 GB in 4-bit. Kaggle's
T4 (16 GB) failed twice — once on the fp32 upcast, once on logits.float(). A10G with
22,5 GB carries it at 15,82 GB peak.
č/ć/š/ž/đ missing from normalisation."kann musik"."ne radi" was required. Check which object a negation negates.["millionen"] was iterated per character, so
"n" matched every answer containing the letter n. The original 23-case set uses a flat
list, new cases use groups; a runner must accept both.And the most expensive one, from the 3B tier: repeat_penalty 1.1 caused a factual
error. Adapter, merged model and the same model with a system prompt all answered
correctly — only the Ollama build with the token penalty failed.
Built by one person, without venture capital, on a laptop with 8 GB of VRAM.
PayPal: https://www.paypal.com/qrcodes/managed/bfb5b6b5-c04c-4836-af60-987dff8cff61
A ⭐ helps as much as a donation.
OpenAmer-Desktop ist die vierte Stufe der OpenAmer-Familie: ein 14,7B-Modell für Desktop-Rechner, das offline läuft und 32k Kontext verarbeitet.
Trainiert mit QLoRA auf gemieteter GPU (A10G), weil 14,7B nicht in 8 GB VRAM passen. Gebaut von OpenAmer in der Schweiz.
Der Agent läuft auch auf dem Telefon — nicht ferngesteuert, sondern im Debian-Userspace mit denselben Werkzeugen und Skills. Launcher-APK, gemessene Version und ehrliche Grenzen:
https://openamer-max.github.io/android.html
Agent-Quellcode: https://github.com/openamer-max/openamer
Für OpenAmer Agent zu wenig: der Agent verlangt mindestens 64.000 Token Kontext. Diese Stufe trägt den Agenten nicht — für Chat und API-Aufrufe ist sie vollständig nutzbar.
Qwen/Qwen2.5-14B-Instruct