Two of the most expensive brains in tech were given the same job: build the Eiffel Tower. Claude Opus 5.5 took 2 hours and ran up a $33.45 bill. Gemini 4 Argon got it done in one shot for $0.23.
That works out to approximately 145 times cheaper for Google’s model.
The Eiffel Tower showdown, by the numbers
What makes the gap interesting is that it can’t be explained by sticker prices alone. Gemini 4 Argon launched with introductory pricing of $2 per million input tokens and $10 per million output tokens. That rate runs until November 30, 2026.
Claude Opus 5.5 lists at $4 per million input tokens and $20 per million output tokens. Anthropic’s pricing represents a 20% decrease from its previous rates. So on a per-token basis, Argon is half the price of Opus 5.5.
Half price does not get you to 145x. The rest of the gap comes from how much work each model did to reach the finish line. A model that nails a task on the first pass burns far fewer tokens than one that iterates for two hours, and tokens are what you pay for.
Beyond the tower: benchmarks and specs
Google DeepMind unveiled Gemini 4 Argon on September 30, 2026. The model became generally available on October 13, 2026. Anthropic had shipped Claude Opus 5.5 just days earlier, on September 22, 2026.
In Google’s report, Argon led with a 68.9% score on the Vals Index. In shared benchmarks between the two models, Argon came out ahead in 14 out of 18 categories. Its strongest showings were in legal and finance applications, and it also posted solid results on agentic coding tests.
Opus 5.5 scored higher on independent aggregates such as the Artificial Analysis Intelligence Index. It also performed well on terminal-heavy benchmarks, which test a model’s ability to operate inside command-line environments.
Argon supports up to 1 million tokens for unified input and output. Opus 5.5 caps output at 128,000 tokens. Argon can produce much longer finished work in a single response, which matters for jobs like drafting lengthy documents or generating large codebases. Opus 5.5 instead emphasizes adaptive reasoning, meaning it adjusts how hard it thinks based on the problem in front of it.
Why a $0.23 tower matters
The Eiffel Tower result reframes the competition around cost per task rather than cost per token. A model with a lower per-token price can still be the expensive option if it needs many more passes to finish the job.
Opus 5.5 already cut its prices by 20%. If a rival can finish certain jobs on the first try, list-price cuts may not close the gap on real workloads.
Argon’s $2 and $10 rates are introductory and expire on November 30, 2026. Any enterprise budgeting around that figure should plan for what happens after the promotion ends.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.