Providers And Models

Providers And Models

Verified against GitMind 5.0.0 on June 7, 2026.

GitMind has 17 built-in providers plus the Pro-only Custom API provider. Cloud providers receive the selected diff and prompt. Provider catalogs change; use Load Available Models and the searchable model picker when available.

Provider	Authentication / setup	Default	Discovery
Google Gemini	API key	gemini-3.1-flash	Yes
Hugging Face	Access token	mistralai/Mistral-7B-Instruct-v0.3	Yes
Ollama	Local server; no key	phi4, http://localhost:11434	Local models
Mistral	API key	mistral-small-4	Yes
Cohere	API key	command-a	Yes
OpenAI	API key	gpt-5.5-instant	Yes
Together AI	API key	meta-llama/Llama-3.3-70B-Instruct-Turbo	Yes
OpenRouter	API key	google/gemma-3-27b-it:free	Yes
Anthropic	API key	claude-sonnet-4.6	Yes
MiniMax	API key	MiniMax-M2.7	Yes
GitHub Copilot	Signed-in Copilot extension/subscription	auto	Detects available models
DeepSeek	API key	deepseek-v4-flash	Yes
xAI Grok	API key	grok-4.3	Yes
Groq	API key	meta-llama/llama-4-scout-17b-16e-instruct	Yes
Perplexity	API key	gpt-5.5-computer	Yes
Z.ai	API key; regular or coding endpoint	glm-5.1, coding endpoint	Yes
NVIDIA hosted NIM	NVIDIA Build API key	meta/llama-3.3-70b-instruct	Yes
Custom API	Pro; endpoint-specific auth	None	User configured

Setup And Checks

1. Select gitmind.apiProvider.
2. Enter the provider key, URL, or authentication required above.
3. Load models, search, and select one.
4. Use Check API Setup to validate connectivity and authentication.
5. Use Check Rate Limits for provider-reported limits. Results vary by provider and the check itself may consume a small request.

Ollama must be running and the selected model must already be pulled. GitHub Copilot must be installed, signed in, and licensed. Z.ai's gitmind.zai.endpoint selects regular or coding.

For private gateways and compatible endpoints, see Custom API Guide. For provider failures, see Troubleshooting And FAQ.