Are GPT-5.2's new powers enough to surpass Gemini 3? Try it and see

Summary

The article discusses the release of OpenAI's GPT-5.2, highlighting its improved capabilities for professional tasks and competition with other AI models. It also mentions safety improvements and access details for users.

Key Takeaways

1. OpenAI released GPT-5.2, its latest model, on Thursday.
2. GPT-5.2 is built for professional tasks and rivals experts.
3. OpenAI noted in the announcement that GPT-5.2 Thinking hallucinates less than 5.1 Thinking by 30%, which it said should encourage enterprise users to worry less about encountering mistakes when using the model for research and analysis.

Statement Breakdown

• Claimed Facts: 60% of statements the article presents as facts
• Opinions: 25% of statements classified as editorial or subjective
• Claims: 15% of statements surfaced for additional reader evaluation

Credibility & Bias Reasoning

Credibility assessment: The article primarily relies on OpenAI's announcements and internal reports, which may present a biased view. ZDNet's reporting adds some balance, but the heavy reliance on company statements lowers overall credibility. The inclusion of a disclosure about a lawsuit against OpenAI is a positive factor.

Bias assessment: Technological Advancement Advocacy. The article frames GPT-5.2's release as a competitive move against Google and Anthropic, highlighting its capabilities and improvements. It emphasizes the economic value and professional applications of the model, suggesting a bias towards promoting technological advancement and OpenAI's products. The focus on positive aspects and competitive advantages indicates a pro-technology stance.

Note: Be cautious of claims made by OpenAI regarding GPT-5.2's performance. Cross-reference information with independent sources to verify accuracy.

Credibility flag: Verify Claims

Claimed Facts (7)

• This is a verifiable fact about the release of the model.
• This is a direct quote from OpenAI's announcement.
• This is a reported performance metric on a specific benchmark.
• This is a statement about performance on industry benchmarks.
• This is a direct statement from OpenAI regarding model deprecation.
• This is a statement about the rollout plan for the model.
• This is a reported reason for the model's fast-tracking.

Opinions (6)

• This is a subjective assessment of the model's capabilities.
• This is a promotional statement about the model's capabilities.
• This is a subjective assessment of the model's safety improvements.
• This is a subjective assessment of the impact of pretraining changes.
• This is a subjective assessment of the model's performance.
• This is an interpretation of OpenAI's motivations.

Claims (5)

• The claim of 11x speed and 1% cost is difficult to verify and may be exaggerated.
• This is a broad claim without specific evidence or independent verification.
• This is a vague and unsubstantiated claim about future developments.
• The 30% reduction in hallucination is difficult to verify and may be misleading.
• The claim of saving an hour each day is based on OpenAI's own study, which may be biased.

Key Sources

• Sabrina Ortiz — Author
• ZDNET — News Source
• OpenAI — Company
• The Information — News Source
• Mark Chen — Chief Research Officer at OpenAI