> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bitscale.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Models

> Run a prompt through GPT, Claude or Gemini models.

Run a prompt through any of these models. Every model takes the same request; pick one by its tool ID.

| Model | Tool ID | Credits per run |
| - | - | - |
| Claude Haiku 4.5 (Anthropic) | `claude-haiku-4-5` | 1 |
| Claude Opus 4.6 (Anthropic) | `claude-opus-4-6` | 4 |
| Claude Sonnet 4.6 (Anthropic) | `claude-sonnet-4-6` | 2.5 |
| Gemini 2.5 Flash (Google) | `gemini-2.5-flash` | 1 |
| Gemini 2.5 Flash Lite (Google) | `gemini-2.5-flash-lite` | 0.5 |
| Gemini 3 Flash Preview (Google) | `gemini-3-flash-preview` | 1 |
| Gemini 3 Pro Preview (Google) | `gemini-3-pro-preview` | 2 |
| Gemini 3.5 Flash Lite (Google) | `gemini-3.5-flash-lite` | 1 |
| Gemini 3.8 Flash (Google) | `gemini-3.8-flash` | 1.5 |
| GPT-4.1 (OpenAI) | `gpt-4.1` | 1 |
| GPT-4o (OpenAI) | `gpt-4o` | 1 |
| GPT-4o Mini (OpenAI) | `gpt-4o-mini` | 0.1 |
| GPT-5 (OpenAI) | `gpt-5` | 2 |
| GPT-5 Mini (OpenAI) | `gpt-5-mini` | 0.5 |
| GPT-5 Nano (OpenAI) | `gpt-5-nano` | 0.1 |
| GPT-5.2 (OpenAI) | `gpt-5.2` | 3 |
| GPT-5.6 Luna (OpenAI) | `gpt-5.6-luna` | 0.25 |
| GPT-5.6 Sol (OpenAI) | `gpt-5.6-sol` | 4 |
| GPT-5.6 Terra (OpenAI) | `gpt-5.6-terra` | 2.5 |

## Options

Optional configuration. Send these in `options`.

| Name | Type | Required | Default | Description |
| - | - | - | - | - |
| `extendedThinking` | boolean | No | `false` | Enable extended/long thinking — the model reasons longer before answering (higher latency/cost). Not every model offers it; check the model's schema. Defaults to false. |
| `webSearch` | boolean | No | `false` | OPT-IN web search. Enabling web search adds latency and cost. |

## Prompt

This tool needs a `prompt`: the full instruction, with the data to work on written into it.

`extendedThinking` is not available on `gpt-4o-mini`, `gpt-5`; sending it returns `400`.

## Output

The answer comes back as text in `result.raw`. Asking for JSON in the prompt still returns a string, often inside a Markdown code fence, so parse it yourself.

```json theme={null}
{
  "tool_id": "claude-haiku-4-5",
  "credit": 1,
  "raw": "Stripe sells payment APIs and financial infrastructure for internet businesses."
}
```

## Example

```bash theme={null}
curl -X POST "https://api.bitscale.ai/api/v1/tools/claude-haiku-4-5/run" \
  -H "X-API-KEY: $BITSCALE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "In one sentence, what does stripe.com sell?"}'
```

See [Tools API](https://docs.bitscale.ai/api-reference/tools/overview.md) for the response format and polling.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.