The world of AI is fiercely competitive. Each year, the main companies go back and forth, all trying to take the top spot as the best and most innovative company, topping leaderboards and becoming the go-to choice of the world.
However, while there is a long list of options to choose from, three have been fighting it out for the top spot. These are OpenAI’s ChatGPT, Claude from Anthropic, and Google’s very own Gemini. With its most recent update, Anthropic might have just become the leading choice in a number of ways. Here’s why.
Coding
Ever since AI chatbots began allowing vibe coding (the ability to write code via prompts), Anthropic has been a leader in the field. Since then, this has become a major focus for the company, pushing noticeable improvements in this area with each iteration.
When it came to the launch of Claude Opus 4.5, this was no different. In Anthropic’s own testing, Opus 4.5 outperformed both Gemini 3 Pro and GPT-5.1 Pro.
This is not to say that its competitors are worse models overall. Gemini 3 has outperformed Claude in its understanding of graduate-level challenges, as well as writing tasks.
However, for Anthropic, this newest model is aimed at being good enough at coding to outperform a human.
When Anthropic interviews engineering candidates, they are given what the company claims is ‘a notoriously difficult take-home exam”. This test is also given to every new model as part of its testing period.
Within the 2-hour time limit that is given for the test, Claude Opus 4.5 scored higher than any human candidate ever. The exam is designed to test performance under pressure, as well as judgement and technical abilities.
For Anthropic, they are building towards a model that can code better than a human, while doing it in half the time.
A workplace leader
When Claude’s latest version of Opus was announced, Anthropic promoted an area of its chatbot that isn’t often spoken about. Gemini and ChatGPT can be used for work and productivity, but Opus 4.5 has been marketed heavily as your workplace tool.
It integrates with Microsoft’s tools, like Word, PowerPoint, and Excel. This is a feature that is available in the likes of ChatGPT, too, and Gemini does the same, but with Google’s suite of tools.
The difference is that Anthropic has made this a core feature, not an add-on. Anthropic wants to become the go-to tool that you use in the workplace.
With this update, Anthropic launched Claude for Excel. This is a feature that understands your entire library of Excel documents and can handle huge chunks of data, while also creating and using detailed, complicated formulas.
This kind of advancement could save people hours at work of time, eliminating the tedious tasks of working through spreadsheets or creating long, detailed formulas to gather information.
Safety focus
Anthropic doesn’t care much for bells and whistles. An area that it has put a lot of attention on, however, is its safety.
Anthropic claims that Claude Opus 4.5 is “the most robustly aligned model we have released to date, and we suspect, the best-aligned frontier model by any developer”.
This refers to Claude’s ability to deal with malicious attacks or attempts to trick the model. Claude Opus 4.5 scored considerably lower on a test of how often it shows concerning behavior compared to both Gemini 3 Pro and GPT-5.1 Pro.
In other words, it is better prepared for prompt injections (where people maliciously hide prompts in text to try and trick an AI model). It is also less likely to fall for malicious attempts to hijack the model in other ways.
As AI has more control in our lives, and is expected to take on tasks on our behalf, an increased safety measure is crucial.
Price concerns
However, despite all of this, the Claude upgrade probably won’t actually be relevant to most people (at least at first). Anthropic has locked Claude Opus 4.5 behind a big pay wall, requiring a $90 a month investment to use it.
Compared to the much lower entry prices for both Gemini 3 and GPT-5.1 (both around $20 a month), that will seem pretty expensive.
Claude Opus 4.5 is very much designed for heavy AI users, pushing their AI chatbot assistant to the extremes. For the average person doing the odd bit of coding, some queries throughout the day and an odd bit of deep research, this will all feel a bit overkill.
But for those who are finding they are reaching for AI throughout the work day, loading it up with complicated tasks and relying on it frequently, Opus 4.5 just might be the best model out there now.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
More from Tom's Guide
- Ads could soon be coming to ChatGPT — here’s what we know
- I just tried the new OpenAI x NORAD holiday tools with my kids — here’s everything ChatGPT can do now
- Apple Intelligence's future uncertain following big leadership changes in Apple's AI team — here's why this could be the best outcome
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
