Providers & Models

Kolega Code is provider-agnostic. You choose which LLM provider to use and which model to use for each of several roles.

Supported providers

The agent runtime recognizes these providers:

Provider	Identifier	API key variable
Anthropic	anthropic	ANTHROPIC_API_KEY
OpenAI	openai	OPENAI_API_KEY
Google	google	GOOGLE_API_KEY
Groq	groq	GROQ_API_KEY
Together.ai	together	TOGETHER_API_KEY
Fireworks.ai	fireworks	FIREWORKS_API_KEY
x.ai (Grok)	xai	XAI_API_KEY
DashScope (Alibaba)	dashscope	DASHSCOPE_API_KEY
Moonshot (Kimi)	moonshot	MOONSHOT_API_KEY
DeepSeek	deepseek	DEEPSEEK_API_KEY
Z.AI (GLM Coding Plan)	zai	ZAI_API_KEY
Kimi Coding Plan	kimi_coding	KIMI_CODING_API_KEY
Llama (local)	llama	none — local model

Tip

Z.AI GLM Coding Plan. The zai provider targets Z.AI’s Anthropic-compatible endpoint (https://api.z.ai/api/anthropic) and exposes GLM-5.2 (default) and GLM-5.1. Set ZAI_API_KEY to your GLM Coding Plan key (the key Z.AI also documents as ANTHROPIC_AUTH_TOKEN).

Tip

Kimi Coding Plan. The kimi_coding provider targets Kimi’s Anthropic-compatible coding endpoint (https://api.kimi.com/coding/), which is separate from the standard Moonshot API used by the moonshot provider. It exposes a single stable model ID, kimi-for-coding, that the backend auto-routes to the latest underlying model: with thinking effort auto (enabled) requests route to K2.7 Code, and with none (disabled) they route to K2.6. Set KIMI_CODING_API_KEY to your Kimi Coding Plan key.

Note

The Settings tab in the TUI currently offers a curated shortlist — Moonshot (Kimi K2.7 Code and Kimi K2.6) and DeepSeek (DeepSeek V4 Pro). The full provider list above is available through environment variables and CLI flags. See Settings & API Keys.

Model roles

Rather than one model for everything, Kolega Code uses three model slots, so you can match cost and capability to the job:

Role	Purpose	Default
Long context	The main coding model — drives the conversation and most work	anthropic / claude-opus-4-7
Fast	Quick, cheap utility calls	anthropic / claude-haiku-4-5-20251001
Thinking	Extended-reasoning (“think hard”) operations	anthropic / claude-opus-4-7 (medium effort)

These built-in role defaults are used only when the corresponding provider or role is explicitly selected without a model. API key variables alone do not make these defaults active.

When you pick a single provider/model in the Settings UI, that selection is applied to all roles. To assign roles individually, use environment variables or CLI flags.

How a model gets chosen

For each role, Kolega Code resolves the provider and model in this order (first match wins):

1. CLI flags — e.g. --provider, --model, --fast-model, …
2. Environment variables — e.g. KOLEGA_CODE_PROVIDER, KOLEGA_CODE_MODEL, the per-role variants, and a project-local .env file.
3. Saved Settings — the active provider/model from the Settings tab.

A configuration is only valid if a provider/model has been selected and every role’s provider has an API key available (the local llama provider needs none). If a key is missing, the CLI reports exactly which environment variable is required.

Set models per role

Override any role from the command line:

kolega-code . \
  --provider anthropic --model claude-opus-4-7 \
  --fast-provider anthropic --fast-model claude-haiku-4-5-20251001 \
  --thinking-provider anthropic --thinking-model claude-opus-4-7 \
  --thinking-effort medium

Or with environment variables — see Environment Variables for the complete list.

Thinking effort

Thinking effort is model-specific. Kolega Code validates the selected value against the active model and resets it to the new model’s default when you switch models in the TUI.

Provider/model	Values	Kolega default
Anthropic claude-opus-4-7	low, medium, high, xhigh, max	medium
DeepSeek deepseek-v4-pro	none, high, max	high
Moonshot kimi-k2.7-code	auto	auto
Moonshot kimi-k2.6	auto, none	auto
Z.AI glm-5.2	high, max	max
Z.AI glm-5.1	auto, none	auto
Google gemini-2.5-pro	auto, low, medium, high	medium
OpenAI reasoning models	low, medium, high	medium

Check what’s active

kolega-code doctor --project .

doctor prints the resolved provider/model for each role (long, fast, thinking) and whether the necessary API keys are present.