# `Gemini.Types.GenerationConfig.ThinkingConfig`
[🔗](https://github.com/nshkrdotcom/gemini_ex/blob/v0.13.0/lib/gemini/types/common/generation_config.ex#L9)

Configuration for thinking/reasoning in Gemini models.

## Gemini 3 (Recommended)

Use `thinking_level` for Gemini 3 models:
- `:minimal` - Minimal thinking (Gemini 3 Flash only)
- `:low` - Minimizes latency and cost. Best for simple tasks.
- `:medium` - Balanced thinking (Gemini 3 Flash only)
- `:high` - Maximizes reasoning depth (default for Gemini 3).

## Model Support

- **Gemini 3 Pro**: `:low`, `:high`
- **Gemini 3 Flash**: `:minimal`, `:low`, `:medium`, `:high`

## Gemini 2.5 (Legacy)

Use `thinking_budget` for Gemini 2.5 models:
- `0` - Disable thinking (Flash/Lite only)
- `-1` - Dynamic thinking
- Positive integer - Fixed token budget

## Important

You cannot use both `thinking_level` and `thinking_budget` in the same request.
Doing so will return a 400 error from the API.

# `t`
[🔗](https://github.com/nshkrdotcom/gemini_ex/blob/v0.13.0/lib/gemini/types/common/generation_config.ex#L44)

```elixir
@type t() :: %Gemini.Types.GenerationConfig.ThinkingConfig{
  include_thoughts: boolean() | nil,
  thinking_budget: integer() | nil,
  thinking_level: thinking_level() | nil
}
```

Thinking configuration for Gemini models

# `thinking_level`
[🔗](https://github.com/nshkrdotcom/gemini_ex/blob/v0.13.0/lib/gemini/types/common/generation_config.ex#L41)

```elixir
@type thinking_level() :: :unspecified | :minimal | :low | :medium | :high
```

---

*Consult [api-reference.md](api-reference.md) for complete listing*