GPT-5 vs. GPT-4o: What Does Ars Technica Say?
Side-by-Side Evaluation
-
Ars Technica tested GPT-5 and GPT-4o on a variety of tasks, including creative writing and technical explanations.
-
GPT-5 edged out with a narrow win: 4 preferred responses vs. 3 for GPT-4o, plus one tie.
(Ars Technica) -
GPT-5 generally produced concise, efficient answers.
-
GPT-4o, however, stood out for its roomier, more expressive tone, offering richer detail and emotional resonance.
(Medial summary)
User Experience & Feedback
-
GPT-5 excels in performance and complexity handling, especially in technical domains.
-
Many users, however, prefer GPT-4o for its warmth and conversational style—attributes highlighted in real-world settings.
(Reddit insights)
Reddit+1Reddit+1
Broader Reception & Benchmarks
-
GPT-5 demonstrates notable gains in reasoning, coding, and multimodal tasks, along with reduced hallucinations and faster responses.
(Tom’s Guide, Business Insider) -
Still, some critics lament the loss of GPT-4o’s personality, prompting OpenAI to restore it for Plus users and introduce customizable GPT-5 personalities.
(OpenAI response)
Business InsiderWindows Central
Final Takeaway
GPT-5 is technically stronger—cleaner answers, better reasoning, and multimodal capability. Yet some users might miss GPT-4o’s engaging style. If your priorities lean toward precision and advanced functions, GPT-5 shines. For emotional resonance, accessibility, and expressivity, GPT-4o still holds appeal.