Kaiden - Models

LLM Providers

Connection status for each provider is shown here. API keys and endpoints are configured on a dedicated page per provider (not a modal).

OpenAI

GPT-4o, GPT-4o-mini, o-series

Connected Verified 2h ago

Credentials accepted; models reachable.

Configure

Anthropic

Claude 3.5 / 4

Connected Verified 1d ago

API responding normally.

Configure

Google AI

Gemini (AI Studio / Vertex)

Not configured

No API key stored. Add one on the configuration page.

Configure

Azure OpenAI

Enterprise deployments

Error Check failed

Could not reach the resource endpoint. Update URL and key on the configuration page.

Configure

Models

Turn models on or off for chat and agents. Only models from connected providers appear here.

Name	Size	Runtime
composer-2-fast Cloud · enabled	—	Composer
gpt-5.3-codex Cloud · enabled	—	OpenAI
gpt-5.4-medium Cloud · enabled	—	OpenAI
claude-4.6-sonnet-medium Cloud · enabled	—	Anthropic
claude-4.6-opus-high Cloud · enabled	—	Anthropic
composer-2 Cloud · disabled	—	Composer
composer-1.5 Cloud · disabled	—	Composer
gpt-5.3-codex-low Cloud · disabled	—	OpenAI

In-house · OpenShift AI

The in-house model source is OpenShift AI only: connect your cluster API (or dashboard URL) and an OpenShift OAuth token. Approved models from your Data Science / serving stack appear below after sync.

OpenShift AI connection

No other in-house catalog types are configured here—only OpenShift AI.

Source OpenShift AI

Cluster API URL

OpenShift token

Use the same API endpoint you would pass to oc login. Token needs access to model serving / OpenShift AI resources in your project.

Name	Size	Age
ibm-granite-3.3-8b-instruct OpenShift AI · Apache-2.0 VRAM est. 8.2 GiB	4.8 GB	2 weeks
mistral-small-internal OpenShift AI VRAM est. 11 GiB	6.1 GB	1 month
llama-3.1-70b-instruct OpenShift AI · pending approval RAM usage: N/A	—	—

Local · Ollama & Ramalama

The extension probes for Ollama and Ramalama on the host: binary or container presence, daemon reachability, then lists models currently loaded (mock detection for this prototype).

Ollama

Installed Running 127.0.0.1:11434

Detected via CLI and HTTP /api/version. Loaded models reflect the inference server state.

Ramalama

Installed Not running Podman service

Start the Ramalama service to expose models; when stopped, no loaded models are reported.

Running on local runtimes

Name	Size	Runtime
llama3.2:3b GGUF · loaded in memory RAM usage: 3.93 GiB	2.0 GB	Ollama
mistral:latest GGUF · loaded in memory RAM usage: 4.4 GiB	4.1 GB	Ollama
qwen2.5:7b Not loaded — Ramalama stopped RAM usage: N/A	4.7 GB	Ramalama