Models
Providers
Neovate Code supports most popular providers and models by default. And you can also add any OpenAI compatible provider. Check out the Providers page to learn more.
Select a model
Once you have configured the providers, you can run /model
to select a model. The selected model will be saved in the global settings in the ~/.neovate/config.json
file.
/model
Also, you can use neovate config
to select a model.
# Set global model
neovate config set model openai/gpt-4o -g
# Set project model
neovate config set model openai/gpt-4o
# Get global model
neovate config get model -g
Recommended models
Here are so many models, and we can’t test all of them. Here are some recommended models which have been tested and are working well.
- Claude Sonnet 4
- Claude Opus 4
- Qwen3 Coder
- Qwen3 Max
- GPT 4.1
- Gemini 2.5 Pro
Specify a model temporarily
You can specify a model temporarily by adding -m, --model
to your command. The format is provider_id/model_id
. It’s priority is higher than the model configured in the settings.
neovate -m openai/gpt-4o "Create a testcase for @src/components/Button.tsx"
Model Aliases
Model aliases are shortcuts for models. You can use them to quickly select a model.
neovate -m 4o
Here are some built-in aliases for convenience:
deepseek
: deepseek/deepseek-chatr1
: deepseek/deepseek-reasoner41
: openai/gpt-4.14
: openai/gpt-44o
: openai/gpt-4oflash-lite
: google/gemini-2.5-flash-liteflash
: google/gemini-2.5-flashgemini
: google/gemini-2.5-progrok
: xai/grok-4grok-code
: xai/grok-code-fast-1sonnet
: anthropic/claude-sonnet-4-20250514sonnet-3.5
: anthropic/claude-3-5-sonnet-20241022sonnet-3.7
: anthropic/claude-3-7-sonnet-20250219sonnet-3.7-thinking
: anthropic/claude-3-7-sonnet-20250219-thinkingk2
: moonshotai-cn/kimi-k2-0711-previewk2-turbo
: moonshotai-cn/kimi-k2-turbo-preview