Alle Modelle sind über eine OpenAI-kompatible API erreichbar:
Base URL: https://llm.ki4buw.de/v1
Beispiel-Request (curl):
curl https://llm.ki4buw.de/v1/chat/completions \
-H "Authorization: Bearer <DEIN_TOKEN>" \
-H "Content-Type: application/json" \
-d '{"model": "<MODELLNAME>", "messages": [{"role":"user","content":"Hallo"}]}'
Den Zugangstoken erhältst du von KI4BUW. Die verfügbaren Modellnamen siehst du in der Tabelle unten.
| Modell | Status | Letzter Check | Latenz | Uptime 24h | Uptime 7d |
|---|---|---|---|---|---|
| Qwen-Image-2.1 |
Offline
The read operation timed out
|
2026-10-27T07:59:05.473078+00:00 | 30041 ms | 0.0 % | 0.0 % |
| bge-reranker-v2-m3 |
Offline
HTTP 400: {"error":{"message":"litellm.BadRequestError: Hosted_vllmException - The model does not support Chat Completions API. Received Model Group=bge-reranker-v2-m3\nAvailable Model Group Fallbacks=None","type":null,"param":null,"code":"400"}}
|
2026-10-27T07:59:05.473078+00:00 | 63 ms | 0.0 % | 0.0 % |
| deepseek-r1:32b | Online | 2026-10-27T07:59:05.473078+00:00 | 6151 ms | 100.0 % | 100.0 % |
| defacto-model |
Offline
HTTP 500: {"error":{"message":"litellm.InternalServerError: InternalServerError: OpenAIException - Connection error.. Received Model Group=defacto-model\nAvailable Model Group Fallbacks=None","type":null,"param":null,"code":"500"}}
|
2026-10-27T07:59:05.473078+00:00 | 14297 ms | 0.0 % | 0.0 % |
| llama3.1:70b | Online | 2026-10-27T07:59:05.473078+00:00 | 11505 ms | 100.0 % | 96.7 % |
| llama4 | Online | 2026-10-27T07:59:05.473078+00:00 | 22248 ms | 100.0 % | 98.4 % |
| ollama/jeffh/intfloat-multilingual-e5-large-instruct:f16 |
Offline
HTTP 500: {"error":{"message":"litellm.APIConnectionError: OllamaException - {\"error\":\"\\\"jeffh/intfloat-multilingual-e5-large-instruct:f16\\\" does not support generate\"}. Received Model Group=ollama/jeffh/intfloat-multilingual-e5-large-instruct:f16\nAvailable Model Group Fallbacks=None","type":null,"pa
|
2026-10-27T07:59:05.473078+00:00 | 4850 ms | 0.0 % | 0.0 % |
| ollama/qwen3.8:27b | Online | 2026-10-27T07:59:05.473078+00:00 | 9456 ms | 100.0 % | 98.4 % |
| openai/qwen3 | Online | 2026-10-27T07:59:05.473078+00:00 | 159 ms | 100.0 % | 95.1 % |
| openai/qwen3.8 | Online | 2026-10-27T07:59:05.473078+00:00 | 391 ms | 100.0 % | 98.4 % |
| oski/inferenz-gpt-oss-120b | Online | 2026-10-27T07:59:05.473078+00:00 | 185 ms | 100.0 % | 96.7 % |
| oski/inferenz-mistral-small-4-119b | Online | 2026-10-27T07:59:05.473078+00:00 | 429 ms | 100.0 % | 98.4 % |
| oski/inferenz-qwen3-8-27b | Online | 2026-10-27T07:59:05.473078+00:00 | 277 ms | 100.0 % | 96.7 % |
| oski/inferenz-qwen3-embedding-8b | Online | 2026-10-27T07:59:05.473078+00:00 | 143 ms | 100.0 % | 98.4 % |
| qwen3-embedding:8b | Online | 2026-10-27T07:59:05.473078+00:00 | 2921 ms | 100.0 % | 98.4 % |
| qwen3.5:27b | Online | 2026-10-27T07:59:05.473078+00:00 | 5846 ms | 100.0 % | 98.4 % |