Added support for gpt-5, gpt-5-mini, and gpt-5-nano Chat Completions models.

Author: veryspry

class-gwiz-gf-openai.php

@@ -261,16 +261,25 @@ public function init() {
 	public function get_openai_models() {
 		$models = array(
 			'chat/completions' => array(
+				'gpt-5'         => array(
+					'description' => __( 'OpenAI\'s flagship model for coding, reasoning, and agentic tasks across domains. Context length: 128k. <a href="https://platform.openai.com/docs/models/gpt-5" target="_blank">More Details</a>', 'gravityforms-openai' ),
+				),
+				'gpt-5-mini'    => array(
+					'description' => __( 'OpenAI\'s faster, more cost-efficient version of GPT-5. It\'s great for well-defined tasks and precise prompts. Context length: 128k. <a href="https://platform.openai.com/docs/models/gpt-5-mini" target="_blank">More Details</a>', 'gravityforms-openai' ),
+				),
+				'gpt-5-nano'    => array(
+					'description' => __( 'OpenAI\'s fastest, cheapest version of GPT-5. It\'s great for summarization and classification tasks. It\'s great for well-defined tasks and precise prompts. Context length: 128k. <a href="https://platform.openai.com/docs/models/gpt-5-mini" target="_blank">More Details</a>', 'gravityforms-openai' ),
+				),
 				'gpt-4o'             => array(
 					'description' => __( 'OpenAI\'s fastest and most affordable flagship model. Context length: 128k. <a href="https://platform.openai.com/docs/models/gpt-4o" target="_blank">More Details</a>', 'gravityforms-openai' ),
 				),
-				'gpt-4-turbo' => array(
+				'gpt-4-turbo'   => array(
 					'description' => __( 'OpenAI\'s previous high-intelligence model. Context length: 128k. <a href="https://platform.openai.com/docs/models/gpt-4-turbo-and-gpt-4" target="_blank">More Details</a>', 'gravityforms-openai' ),
 				),
-				'gpt-4o-mini' => array(
+				'gpt-4o-mini'   => array(
 					'description' => __( 'OpenAI\'s most cost-efficient small model. Context length: 128k. <a href="https://platform.openai.com/docs/models/gpt-4o-mini" target="_blank">More Details</a>', 'gravityforms-openai' ),
 				),
-				'gpt-3.5-turbo'       => array(
+				'gpt-3.5-turbo' => array(
 					'description' => __( 'Inexpensive model for simple tasks. Context length: 16k. <a href="https://platform.openai.com/docs/models/gpt-3-5-turbo" target="_blank">More Details</a>', 'gravityforms-openai' ),
 				),
 			),
@@ -1174,7 +1183,13 @@ public function get_text_from_response( $response ) {
 			return trim( rgars( $response, 'choices/0/message/content' ) );
 		}
 
-		return trim( rgar( $response, 'text' ) );
+		$text = rgar( $response, 'text', '' );
+		if ( ! empty( $response ) ) {
+			// safetly check as trim() doesn't accept null values
+			$text = trim( $text );
+		}
+
+		return $text;
 	}
 
 	/**
@@ -1373,11 +1388,36 @@ public function make_request( $endpoint, $body, $feed ) {
 
 		switch ( $endpoint ) {
 			case 'chat/completions':
-				$body['max_tokens']        = (float) rgar( $feed['meta'], $endpoint . '_' . 'max_tokens', $this->default_settings['chat/completions']['max_tokens'] );
-				$body['temperature']       = (float) rgar( $feed['meta'], $endpoint . '_' . 'temperature', $this->default_settings['chat/completions']['temperature'] );
-				$body['top_p']             = (float) rgar( $feed['meta'], $endpoint . '_' . 'top_p', $this->default_settings['chat/completions']['top_p'] );
-				$body['frequency_penalty'] = (float) rgar( $feed['meta'], $endpoint . '_' . 'frequency_penalty', $this->default_settings['chat/completions']['frequency_penalty'] );
-				$body['presence_penalty']  = (float) rgar( $feed['meta'], $endpoint . '_' . 'presence_penalty', $this->default_settings['chat/completions']['presence_penalty'] );
+				$body['max_completion_tokens'] = (float) rgar(
+					$feed['meta'],
+					$endpoint . '_max_tokens',
+					$this->default_settings['chat/completions']['max_tokens']
+				);
+
+				// temperature is deprecated in certain models, so only set if the user has set an explicit value.
+				$temperature = rgar( $feed['meta'], $endpoint . '_temperature' );
+				if ( $temperature !== null && $temperature !== '' ) {
+					$body['temperature'] = (float) $temperature;
+				}
+
+				// top_p is deprecated in certain models, so only set if the user has set an explicit value.
+				$top_p = rgar( $feed['meta'], $endpoint . '_top_p' );
+				if ( $top_p !== null && $top_p !== '' ) {
+					$body['top_p'] = (float) $top_p;
+				}
+
+				// frequency_penalty is deprecated in certain models, so only set if the user has set an explicit value.
+				$frequency_penalty = rgar( $feed['meta'], $endpoint . '_frequency_penalty' );
+				if ( $frequency_penalty !== null && $frequency_penalty !== '' ) {
+					$body['frequency_penalty'] = (float) $frequency_penalty;
+				}
+
+				// presence_penalty is deprecated in certain models, so only set if the user has set an explicit value.
+				$presence_penalty = rgar( $feed['meta'], $endpoint . '_presence_penalty' );
+				if ( $presence_penalty !== null && $presence_penalty !== '' ) {
+					$body['presence_penalty'] = (float) $presence_penalty;
+				}
+
 				break;
 		}