Is Google’s new model catching up with OpenAI, Anthropic at the frontier?

This report comes from this week’s The Tech Download newsletter. Do you like what you see? You can subscribe here.

2026 was an eventful year for Google AI hopes and dreams.

Things started off well. Gemini 3, launching towards the end of 2025, has placed Google in the AI ​​frontier debate.

But for most of the rest of the year, it was Anthropic and OpenAI that were at the forefront of AI advancements.

Google is undoubtedly hoping that the release of a new flagship model, Gemini 4 Argon, changes that.

When Demis Hassabis resigned as CEO of DeepMind, Google’s AI lab, in August, I wrote that new boss Koray Kavukcuoglu was inheriting a race to catch up to the companies of Sam Altman and Dario Amodei.

Google now touts the model as outperforming top offerings OpenAI and Anthropic in some benchmarks.

The Artificial Analytics Intelligence Index, a composite benchmark score, places Gemini 4 behind only Claude Opus 5.5 and Claude Sonnet 5.5 in its rankings.

“Google Gemini 4 Argon demonstrates advanced reasoning on critical tasks, against benchmarks, particularly legal reasoning, finance and other aspects of enterprise knowledge work, including long-running tasks,” Tim Law, research director for AI at IDC., told me.

Lian Jye Su, chief analyst at Omdia, said that although Google was “late to the cybersecurity and coding game,” Gemini 4 took the company to the frontier of AI, particularly in cybersecurity.

Google was keen to address this point.

The company plans to launch the new model in stages, starting with trusted cybersecurity partners while working with the U.S. government on preliminary security assessments.

“Starting this deployment in this way gives us more confidence, but also allows us to get a trained, robust cyber defense model into the hands of defenders as soon as possible,” Tulsee Doshi, product manager of Google’s Gemini model, told CNBC.

The numerous security incidents caused by malicious AI systems created by OpenAI and Anthropic have created an opportunity for Google to “claim its position as a provider of trust models for safe and secure AI operations,” Su said.

But even if Argon has made Google competitive again in AI, said Nick Patience, head of AI at Futurum Group, the model doesn’t make it the “leader.”

“At a minimum, however, Gemini 4 Argon opens another competitive front for enterprise knowledge work with other leading pioneer models,” Law said.

Google said in its statement that its new model is already being used internally to optimize memory in Google’s data centers, freeing up hundreds of terabytes of memory without purchasing additional hardware. Quantum computing researchers have also used this model.

But the real test, as is always the case with new models, is when they see the light of day and are launched more widely. The company has given no indication of when this might happen.

“While Google touts the success of internal use cases, the final proof will be in enterprise production environments once the model is fully released,” Law said.

OpenAI has abandoned the release of an upcoming model after determining that it did not adequately meet the company’s safety standards.

EspaceX launched Google AI chips into orbit in a move towards space data centers.

OpenAI was sued by a non-profit organization for cyberattacking its models against the startup Hugging Face in July.

California Governor Gavin Newsom bans AI ‘bot bosses’ in landmark state lawoverriding his earlier veto.

Former LinkedIn boss Roslansky set to leave Microsoft, following other departures of leaders.

US President Donald Trump’s renaming of AI to SI (super intelligence for the uninitiated) had an unexpected (although perhaps predictable) consequence.

This led to an “unprecedented” increase in registrations for the small European country where Slovenia is located, which is, you guessed it, .yes.

A spokesperson for Register.si, Slovenia’s official national registry, told the BBC that 44,000 .si web addresses were created in September, up from fewer than 2,000 in August.

This is not the first time that AI has sparked a mad rush into a small country’s domain.

Anguilla, a Caribbean island with a population of fewer than 20,000, has also seen a boom in website listings in recent years.

Its domain: .ai.