Models

List the available models and target them via the model slug.

You select the AI model via the model field of each request. You pass a Kasimir slug (not the provider’s model name). Which models are available to your key depends on your organization's settings — GET /v1/models is always authoritative.

Available models

Slug	Model	Type	Context	Vision	Processing
`gemma-4-31b`	Gemma 4 (31B)	Chat	262k	yes	EU · self-hosted (default)
`gpt-5.6-luna`	GPT-5.6 Luna	Chat	922k	yes	EU · Azure AI Foundry
`gpt-5.6-terra`	GPT-5.6 Terra	Chat	922k	yes	EU · Azure AI Foundry
`gpt-5.6-sol`	GPT-5.6 Sol	Chat	922k	yes	EU · Azure AI Foundry
`gpt-5.4`	GPT-5.4	Chat	272k	yes	EU · Azure AI Foundry
`claude-sonnet-4-6`	Claude Sonnet 4.6	Chat	1M	yes	EU · Anthropic (EU processing)
`mistral-large-3`	Mistral Large 3	Chat	256k	–	EU · Mistral
`grok-4.3`	Grok 4.3	Chat	256k	yes	EU · Azure AI Foundry
`deepseek-v4-pro`	DeepSeek V4 Pro	Chat	131k	–	EU · Azure AI Foundry
`jina-embeddings-v3`	Jina Embeddings v3 (1024d)	Embeddings (1024 dim.)	8k	–	EU · self-hosted

This is the current default catalogue. Your organization may disable models or enable additional (incl. non-EU) models. The Basic plan only includes the self-hosted models gemma-4-31b and jina-embeddings-v3. GET /v1/models always returns the list valid for your key.

Selecting a model

Set the slug in the request's model field:

from openai import OpenAI
client = OpenAI(base_url="https://kasimir.ai/api/v1", api_key="kasi_…")
client.chat.completions.create(model="grok-4.3", messages=[{"role": "user", "content": "Hello"}])

Only vision models (e.g. grok-4.3, gemma-4-31b) process images. For embeddings use jina-embeddings-v3 on the /v1/embeddings endpoint.

Listing models

GET /v1/models returns the models enabled for your key in OpenAI format:

curl https://kasimir.ai/api/v1/models -H "Authorization: Bearer kasi_…"

{ "object": "list", "data": [ { "id": "mistral-large-3", "object": "model", "owned_by": "kasimir" } ] }

Governance

The model must be active for your company and of the right type (chat / embedding).
Models disabled at the workspace level are not callable → 403.
Non-EU models are only available if your company has enabled them → otherwise 403.
On the Basic plan only the self-hosted models are callable; all others are missing from GET /v1/models and return 403.
The response always returns the Kasimir slug in the model field.

BackAPI: Embeddings NextAPI: Errors, Status Codes & Limits