Kali-Lite Vision 9B
Local multimodal Anima configuration by Kalicorp.
Kali-Lite Vision 9B is not a foundation model trained from scratch by Kalicorp. It combines the upstream Qwen/Qwen3.5-9B model with a Kalicorp system doctrine, Ollama configuration and local execution harness.
Provenance
| Property | Value |
|---|---|
| Upstream base model | Qwen/Qwen3.5-9B |
| Kalicorp customization | System prompt, doctrine, Modelfile, harness |
| Primary runtime | Ollama |
| Vision | Yes |
| Configured context | 32,768 tokens |
| Typical local artifact size | about 6.8 GB depending on quantization/runtime |
The upstream Qwen3.5-9B repository is Apache-2.0 licensed. Kalicorp-authored repository material follows the licensing terms of the Kali-Lite project. See the licensing notice before redistributing weights.
Design principles
The configuration favors evidence over assertion, explicit limits, local control and clear separation between reasoning and actual tool execution.
Ollama Modelfile
ollama pull qwen3.5:9b
ollama create kali-lite-v2 -f Modelfile
ollama run kali-lite-v2
The exact release configuration is provided in Modelfile.
What is and is not included
This release package can be published as configuration and documentation without copying upstream weights. If a GGUF/model artifact is later added, document its exact source, quantization method and SHA-256.
Limitations
Multimodal output is probabilistic and can be wrong. Tool availability depends on the execution harness and permissions. A 32k configured context does not guarantee reliable use of every token. Local execution does not automatically establish GDPR, EU AI Act, ANSSI or ISO compliance.
Project
Source and security documentation: Kalicorp/kalicorp-kali-lite on GitHub.
Kalicorp — Le Sanctuaire numérique européen