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

# AI Thinking Fields

> Production-revalidated guide to the thinking and reasoning fields exposed by different Crazyrouter protocols

> Дата обновления: 2026-06-06

# AI Thinking Fields

This page only lists thinking and reasoning surfaces that were revalidated against Crazyrouter production on `2026-03-22`.

The key point is that different protocols expose "thinking" differently:

* OpenAI Responses returns a standalone `reasoning` item
* Claude native Messages returns a `thinking` block
* Gemini Native is most reliably observed through `usageMetadata.thoughtsTokenCount`

## Currently verified fields

| Model             | Protocol           | Verified field or marker           | Notes                                                        |
| ----------------- | ------------------ | ---------------------------------- | ------------------------------------------------------------ |
| `gpt-5.5`         | Responses          | `output[].type = "reasoning"`      | Can return displayable summaries when `summary` is requested |
| `claude-opus-4-8` | Anthropic Messages | `content[].type = "thinking"`      | Returned alongside a `text` block                            |
| `gemini-3.1-pro`  | Gemini Native      | `usageMetadata.thoughtsTokenCount` | Confirms thinking tokens were actually used                  |

<Note>
  In the current recheck, `gpt-5.5` Chat Completions did not reliably return usable `message.reasoning_content`, so it is not treated as the primary observable field here.
</Note>

***

## GPT: Responses `reasoning` item

```json theme={null}
{
  "output": [
    {
      "id": "rs_xxx",
      "type": "reasoning",
      "encrypted_content": "...",
      "summary": [
        {
          "type": "summary_text",
          "text": "..."
        }
      ]
    },
    {
      "type": "message",
      "content": [
        {
          "type": "output_text",
          "text": "Final answer"
        }
      ]
    }
  ]
}
```

Extraction example:

```python theme={null}
response = client.responses.create(
    model="gpt-5.5",
    input="Which is larger, 9.11 or 9.9?",
    reasoning={"effort": "high", "summary": "detailed"}
)

for item in response.output:
    if item.type == "reasoning":
        for part in item.summary:
            if part.type == "summary_text":
                print("Thinking summary:", part.text)
```

***

## Claude: `thinking` block

```json theme={null}
{
  "content": [
    {
      "type": "thinking",
      "thinking": "..."
    },
    {
      "type": "text",
      "text": "Final answer"
    }
  ]
}
```

Extraction example:

```python theme={null}
message = client.messages.create(
    model="claude-opus-4-8",
    max_tokens=320,
    thinking={
        "type": "enabled",
        "budget_tokens": 128
    },
    messages=[
        {"role": "user", "content": "Which is larger, 9.11 or 9.9?"}
    ]
)

for block in message.content:
    if block.type == "thinking":
        print("Thinking:", block.thinking)
    elif block.type == "text":
        print("Answer:", block.text)
```

***

## Gemini: `thoughtsTokenCount`

In the current production recheck, the most reliable observable signal was not a visible reasoning paragraph in the body, but a usage field:

```json theme={null}
{
  "usageMetadata": {
    "thoughtsTokenCount": 120
  }
}
```

This indicates that:

* the thinking budget was actually used
* the request consumed thinking tokens during generation

Extraction example:

```python theme={null}
data = response.json()
thoughts = data.get("usageMetadata", {}).get("thoughtsTokenCount", 0)
print("Thinking tokens:", thoughts)
```

***

## Usage notes

* Do not assume every model returns raw chain-of-thought text
* Do not mix thinking fields across protocols
* For logging, auditing, or UI display, decide per protocol whether to read `reasoning`, `thinking`, or `thoughtsTokenCount`

Related pages:

* [GPT-5 Thinking Mode](/ru/chat/responses/gpt5-thinking)
* [Claude Native Messages](/ru/chat/anthropic/messages)
* [Gemini Native API](/ru/chat/gemini/native)