Thinking

Extended thinking lets Claude reason through complex problems before answering. Useful for math, logic, analysis, and code generation.

The recommended approach for Claude Opus 4.6 and Sonnet 4.6. Claude decides how much to think based on the request:

use Anthropic\Laravel\Facades\Anthropic;

$response = Anthropic::messages()->create([
    'model' => 'claude-opus-4-6',
    'max_tokens' => 16000,
    'thinking' => [
        'type' => 'adaptive',
    ],
    'messages' => [
        ['role' => 'user', 'content' => 'Explain why the sum of two even numbers is always even.'],
    ],
]);

The response content may include thinking blocks before the final text block:

foreach ($response->content as $block) {
    if ($block->type === 'thinking') {
        $block->thinking;  // 'Let me analyze this step by step...'
        $block->signature; // 'WaUjzkypQ2mUEVM36O2TxuC06KN8xyfbJwyem...'
    }

    if ($block->type === 'text') {
        $block->text; // 'Based on my analysis...'
    }
}

Control how much Claude thinks with output_config.effort:

$response = Anthropic::messages()->create([
    'model' => 'claude-opus-4-6',
    'max_tokens' => 16000,
    'thinking' => ['type' => 'adaptive'],
    'output_config' => [
        'effort' => 'medium',
    ],
    'messages' => [...],
]);

For older models (Sonnet 3.7, Opus 4.5, Sonnet 4.5), use a fixed token budget. This is deprecated on Opus 4.6 and Sonnet 4.6:

$response = Anthropic::messages()->create([
    'model' => 'claude-sonnet-4-5',
    'max_tokens' => 16000,
    'thinking' => [
        'type' => 'enabled',
        'budget_tokens' => 10000,
    ],
    'messages' => [...],
]);

Thinking requests often take 30+ seconds. Dispatch them to a queue, and bump your timeout:

ANTHROPIC_REQUEST_TIMEOUT=120

class AnalyzeJob implements ShouldQueue
{
    public $timeout = 150;

    public function handle(): void
    {
        $response = Anthropic::messages()->create([
            'model' => 'claude-opus-4-6',
            'max_tokens' => 16000,
            'thinking' => ['type' => 'adaptive'],
            'output_config' => ['effort' => 'max'],
            'messages' => $this->messages,
        ]);

        // Save the final answer (skip thinking blocks for display)
        $text = collect($response->content)
            ->firstWhere('type', 'text')
            ->text;

        $this->result->update(['content' => $text]);
    }
}

Pass the full content array (thinking blocks included) back in follow-up requests so Claude can verify its previous reasoning:

$followUp = Anthropic::messages()->create([
    'model' => 'claude-opus-4-6',
    'max_tokens' => 16000,
    'thinking' => ['type' => 'adaptive'],
    'messages' => [
        ['role' => 'user', 'content' => 'Solve x^2 + 5x + 6 = 0'],
        ['role' => 'assistant', 'content' => $response->toArray()['content']],
        ['role' => 'user', 'content' => 'Now verify by substitution.'],
    ],
]);

The signature field on each thinking block is used by the server to verify integrity, so don't modify it.

For display options (summarized vs omitted), redacted_thinking blocks, interleaved thinking with tools, and streaming behavior, see the Thinking page in the PHP docs or the Extended thinking guide and Adaptive thinking guide on the Anthropic docs.