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

# Claude Haiku 4.5 Thinking

> Anthropic's Claude Haiku 4.5 with Extended Thinking — fast reasoning at low cost, great for math, coding, and analysis.

Claude Haiku 4.5 Thinking is the Extended Thinking version of Claude Haiku 4.5, available through Starrise AI via OpenAI-compatible interface. Combining Haiku's speed and cost advantages with step-by-step internal reasoning enabled by default, it significantly improves accuracy on complex tasks while remaining far faster and more affordable than large thinking models.

## Key Capabilities

* **OpenAI-compatible** — Drop-in replacement for the OpenAI SDK, no other code changes needed
* **Extended Thinking** — Step-by-step internal reasoning enabled by default
* **Fast and affordable** — Thinking capability at Haiku-level speed and cost
* **Streaming** — Real-time token streaming via SSE
* **Strong reasoning** — Higher accuracy on math, logic, and multi-step problems

## Quick Example

```bash cURL theme={null}
curl https://dayu.cdn-fly.com/v1/messages \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-haiku-4-5-20251001-thinking",
    "messages": [
      { "role": "user", "content": "Solve step by step: A train travels 120 km in 2 hours, then slows down and travels 80 km in 3 hours. What is the average speed for the entire journey?" }
    ]
  }'
```

```python Python theme={null}
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://dayu.cdn-fly.com/v1"
)

response = client.chat.completions.create(
    model="claude-haiku-4-5-20251001-thinking",
    messages=[
        {"role": "user", "content": "Solve step by step: A train travels 120 km in 2 hours, then slows down and travels 80 km in 3 hours. What is the average speed for the entire journey?"}
    ]
)

print(response.choices[0].message.content)
```

```python Streaming theme={null}
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://dayu.cdn-fly.com/v1"
)

stream = client.chat.completions.create(
    model="claude-haiku-4-5-20251001-thinking",
    messages=[
        {"role": "user", "content": "Prove that there are infinitely many prime numbers."}
    ],
    stream=True
)

for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")
```

## Parameters

| Parameter | Type | Required | Description |
| - | - | - | - |
| `model` | string | Yes | Fixed value: `claude-haiku-4-5-20251001-thinking` |
| `messages` | array | Yes | Array of `{ role, content }` objects |
| `max_tokens` | integer | No | Maximum number of tokens to generate |
| `temperature` | float | No | `0`–`2`, controls randomness, default `1` |
| `stream` | boolean | No | Enable SSE streaming, default `false` |
| `top_p` | float | No | Nucleus sampling threshold, default `1` |
| `stop` | string / array | No | Sequences where generation stops |

> **[API Reference](/en/api-reference/model-api/anthropic/claude-haiku-4-5-20251001-thinking)**
>
> View the interactive API Playground for Claude Haiku 4.5 Thinking.


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