Models Library
AgentLens supports 32 models from the Ollama Cloud API, spanning from lightweight 3B models to frontier 1T-parameter models. Each model can be assigned to any pipeline role from the UI dropdown.
| Name | Vendor | Parameters | Size | Capabilities |
|---|---|---|---|---|
gemini-3-flash-previewFallback | – | – | ||
ministral-3:3b | Mistral | 3B | 4 GB | |
gemma3:4b | 4B | 8 GB | ||
ministral-3:8b | Mistral | 8B | 10 GB | |
rnj-1:8b | 8B | 15 GB | ||
gemma3:12b | 12B | 22 GB | ||
ministral-3:14b | Mistral | 14B | 15 GB | |
gpt-oss:20b | OpenAI | 21B | 13 GB | |
devstral-small-2:24b | Mistral | 24B | 48 GB | |
gemma3:27b | 27B | 51 GB | ||
nemotron-3-nano:30bReAct | NVIDIA | 30B | 30 GB | |
qwen3-coder-next | Alibaba | 80B | 76 GB | |
qwen3-next:80bGrader | Alibaba | 80B | 76 GB | |
gpt-oss:120bJudge | OpenAI | 117B | 61 GB |
Models with a colored dot and role badge are the current pipeline defaults: ReAct, Grader, Judge, and Fallback. See Models Overview for the override priority chain. Sizes shown are file sizes on disk.