All models available on IIO AI Platform — updated 2026-05
Fast, capable 7B parameter model. Best for summaries, Q&A, classification, and everyday tasks. Low latency.
High-capability 72B model for complex reasoning, analysis, and long-form generation. Use when 7B is not enough.
Open-source GPT-class model fine-tuned for instruction following. Good balance of speed and quality.
Qwen3 next-gen model with improved reasoning and multilingual capabilities. Excellent for German-language tasks.
High-quality text embeddings for semantic search and RAG. 768-dim vectors. Best general-purpose embedding model.
Qwen3 embedding model with 4096-dim vectors. Better multilingual coverage than nomic-embed.
| Task | Recommended Model | Why |
|---|---|---|
| Simple Q&A, summaries | qwen2.5:7b | Fast, low cost, good enough for most tasks |
| Complex analysis, legal | qwen2.5:72b | Higher reasoning quality, handles nuance |
| German-language tasks | qwen3:8b | Best German comprehension |
| RAG / document search | nomic-embed-text + any chat model | Fast embeddings, proven RAG quality |
| Production chatbot | qwen2.5:7b with fallback to :72b | Speed + quality fallback pattern |
| Cell | Tenant | Models | Status |
|---|---|---|---|
| inhzgx1 | intelego | qwen2.5:7b, nomic-embed-text | ● live |
| inhzgx2 | bilz | qwen2.5:7b, gpt-oss:20b, qwen3-embedding:8b, nomic-embed-text | ● live |
| inhzgx3 | netplans | qwen2.5:7b, nomic-embed-text | ● live |
| inhzgx4 | occ | qwen2.5:7b, nomic-embed-text | ● live |
| inhzgx5 | pm24 | qwen2.5:7b | ● live |
For Developer API access, requests are routed to the nearest available cell automatically.