I tested GPT-5.4, and the answers were really good - just not always what I asked

Summary

The author reviews GPT-5.4 Thinking, highlighting its improved reasoning and analysis capabilities. The review covers image generation, travel planning, and social media analysis, noting both strengths and weaknesses.

Key Takeaways

1. GPT-5.4 Thinking delivers deeper analysis than earlier ChatGPT models.
2. It has strong reasoning, but it sometimes answers questions you didn't ask.
3. Formatting and image generation lag behind the text quality.

Statement Breakdown

• Claimed Facts: 60% of statements the article presents as facts
• Opinions: 30% of statements classified as editorial or subjective
• Claims: 10% of statements surfaced for additional reader evaluation

Credibility & Bias Reasoning

Credibility assessment: The article is a first-person review of GPT-5.4 Thinking, detailing the author's experiences with the AI model. The author provides specific examples and test cases, enhancing credibility. However, the subjective nature of the review and the lack of external validation slightly lower the score.

Bias assessment: Enthusiastic Tech Tester. The author is generally positive about the GPT-5.4 Thinking model, highlighting its strengths while acknowledging its weaknesses. The bias stems from a tech enthusiast's perspective, eager to explore and showcase the capabilities of new AI technologies. The author's focus is on practical application and user experience.

Note: This article presents a subjective review of GPT-5.4 Thinking. Interpret findings as one person's experience, not a definitive evaluation.

Credibility flag: Experiential Assessment

Claimed Facts (7)

• This is a factual statement identifying the name of the AI model being reviewed.
• This is a factual statement about the availability of the AI model.
• This is a factual statement about the author's methodology.
• This is a factual statement about the author's testing process.
• This is a factual statement about the author's experience.
• This is a factual statement about the author's experience.
• This is a factual statement about the AI's response.

Opinions (6)

• This is a subjective opinion about specific aspects of the AI model.
• This is a subjective assessment of the AI model's utility.
• This is a subjective assessment of the AI's image generation capabilities.
• This is a subjective assessment of the AI's travel planning suggestions.
• This is a subjective complaint about the AI's formatting.
• This is a subjective expression of appreciation for the AI's capabilities.

Claims (6)

• This statement lacks specific context and source data, making it difficult to verify.
• This is an unsubstantiated claim about the AI's underlying technology.
• This is an emotional appeal and lacks concrete evidence.
• This is a subjective and potentially exaggerated observation.
• This statement is vague and lacks specific details about the prompt.
• This statement is vague and lacks specific details about the prompt.

Key Sources

• David Gewirtz — Author
• ZDNET — Media
• OpenAI — Company