Model Configuration

N.E.K.O. uses different AI models for different tasks. Each can be individually configured.

Model roles

The model for each role is resolved from the selected assist provider (config/api_providers.json → assist_api_providers), so there is no single global default — each provider ships its own per-role models. The columns below show the shipped defaults for the OpenAI, Claude (Anthropic), and Qwen providers. (The DEFAULT_*_MODEL constants in config/__init__.py are last-resort fallbacks only.)

Role	Config field	OpenAI	Claude	Qwen
Conversation	conversation_model	gpt-5-chat-latest	claude-sonnet-4-6	qwen3.7-plus-2026-05-26
Summary	summary_model	gpt-5-chat-latest	claude-sonnet-4-6	qwen3.7-plus-2026-05-26
Correction	correction_model	gpt-5-chat-latest	claude-sonnet-4-6	qwen3.7-plus
Emotion	emotion_model	gpt-4.1-nano	claude-haiku-4-5-20251001	qwen3.6-flash
Vision	vision_model	gpt-5-chat-latest	claude-sonnet-4-6	qwen3.7-plus-2026-05-26
Agent	agent_model	glm-5v-turbo	claude-opus-4-6	qwen3.7-plus

Note: agent_model is the model used for agent (Computer Use) task grounding. Several providers intentionally ship a strong vision/grounding model here regardless of their own family — e.g. the OpenAI provider's agent_model is glm-5v-turbo (a GLM vision model). These values are copied verbatim from config/api_providers.json and are by design, not typos.

Custom model endpoints

Each model role can use a custom API endpoint. This is configured in core_config.json or via the Web UI:

{
  "conversationModel": "custom-model-name",
  "conversationModelUrl": "https://custom-api.example.com/v1",
  "conversationModelApiKey": "sk-xxxxx"
}

When a custom URL/key is set, it overrides the global assist API provider for that specific role.

Computer Use models

Computer Use requires two vision models:

Role	Default	Purpose
Planning model	qwen3-vl-plus-2025-09-23	Analyze screenshots and plan actions
Grounding model	qwen3-vl-plus-2025-09-23	Locate UI elements for clicking

Configure via core_config.json:

{
  "computerUseModel": "qwen3-vl-plus-2025-09-23",
  "computerUseModelUrl": "https://dashscope.aliyuncs.com/compatible-mode/v1",
  "computerUseModelApiKey": "sk-xxxxx",
  "computerUseGroundModel": "qwen3-vl-plus-2025-09-23",
  "computerUseGroundUrl": "https://dashscope.aliyuncs.com/compatible-mode/v1",
  "computerUseGroundApiKey": "sk-xxxxx"
}

Thinking mode configuration

Some models support "thinking" or "extended reasoning" modes. N.E.K.O. disables these by default for faster responses. The disable format varies by provider:

Provider	Disable format
Qwen, Step, DeepSeek	{"enable_thinking": false}
GLM	{"thinking": {"type": "disabled"}}
Gemini 2.x	{"thinking_config": {"thinking_budget": 0}}
Gemini 3.x	{"thinking_config": {"thinking_level": "low"}}

This is handled automatically in config/__init__.py based on the model name.

Image rate limiting

Setting	Default	Description
NATIVE_IMAGE_MIN_INTERVAL	1.5s	Minimum interval between screen captures
IMAGE_IDLE_RATE_MULTIPLIER	5x	Multiplier when no voice activity