GPT 5s New Powers And Where Competitors Excel

Last Thursday, OpenAI launched the latest version of its hyper-popular AI chatbot, ChatGPT. Sam Altman, OpenAI’s CEO, made some pretty big claims about GPT-5 during its unveiling, comparing the new model's expertise to that of a Ph.D. level expert available on demand.

OpenAI's Hype vs. User Reality

While OpenAI insists GPT-5 is its best AI yet—touting more expertise, greater capability, and fewer mistakes—users aren't so convinced. Many were initially frustrated when OpenAI briefly hid older models like GPT-4o, a change the company has since walked back.

Beyond that, users have complained about GPT-5’s “cold” or “blunt” conversational style, with some comparing it to an “overworked secretary.” Others have noted its stricter safety and refusal policies, pointing out that rivals like Anthropic’s Claude, Google’s Gemini, or Meta’s Llama remain better choices for specific tasks like live coding or custom deployments.

If you find yourself underwhelmed, let's break down the distinct advantages GPT-5 offers, as well as where its competitors still shine.

Key Upgrades: What Can GPT-5 Do?

1. Unified, adaptive reasoning OpenAI highlighted that previous ChatGPT versions made users choose models for different tasks, balancing speed and intelligence. The main innovation of GPT-5 is eliminating this choice. OpenAI claims it's the best all-around option for both quick answers and deep thinking, with particular strengths in coding and math.

2. More accurate answers with fewer hallucinations GPT-5 is said to excel at extended reasoning, with significantly reduced hallucination rates. For instance, HealthBench scores reportedly show up to 80% fewer factual errors in complex scenarios compared to previous models.

3. Rich personalization GPT-5 introduces customizable personalities (like Cynic, Robot, Listener) and improved voice features, allowing users to tailor the chatbot's tone and style without complex prompt engineering. The voice model sounds more natural, and memory enhancements help it learn about users over time. Soon, it will gain access to Gmail and Google Calendar, making it a more useful and personal assistant.

4. Workflow Integration With direct integration for apps like Gmail and Google Calendar, GPT-5 can manage schedules and emails. Businesses can also customize it with their own data. OpenAI calls it “the best coding model on the market” and says it can work autonomously on projects for extended periods, using a self-improvement loop to iterate on its own code.

The Competition: Where Other Chatbots Shine

Despite its power, GPT-5 isn't the only player in the field.

• Claude Opus 4.1 from Anthropic is often considered superior for sustained autonomous coding. Features like “Claude Code” offer specialized developer tools that GPT-5 currently lacks. Claude also projects code visualizations, or “artifacts,” live in its IDE, improving clarity and persistence across sessions.

• Google’s Gemini 2.5 Pro stands out for its seamless video analysis, an area still developing for GPT-5. Gemini’s multimodal engine natively handles short videos, diagrams, and complex visual reasoning. Its connection to Google Search also makes it excellent for accessing real-time data and breaking news, something GPT-5 struggles with.

• Llama 3 from Meta is the top choice for research and privacy-conscious organizations. As an open-source model, it gives developers full control over customization, security, and deployment. This allows for faster experimentation and the ability to run it on private infrastructure, a key advantage over GPT-5's closed API approach.

The Takeaway: An Ongoing AI Arms Race

Even with all the improvements in ChatGPT-5, the AI arms race is far from over. Anthropic, Google, and Meta are keeping pace, particularly in specialized areas like customization, coding, and real-time information. As always in the world of AI, the best approach is to experiment with different models to discover which one best fits your unique workflow and needs.