gemini 4 carbon: google's strongest model, and almost top for a third of the price
launched to cyber defenders first. a week later a work benchmark put it right behind the most expensive model at roughly a third of the cost per task.
what google launched
google announced gemini 4 carbon and calls it its strongest model, with a focus on cybersecurity and long tasks.
google says carbon sits at 59 and holds the top spot on artificial analysis' new intelligence index. the verge reports it sits ahead of claude opus 5.5 and gpt-6 astra there.
google says internal teams have seen it run for days on complex, multi-step engineering work.
who can use it
almost nobody yet.
google is giving it to a subset of cyber defenders first, then gradually to trusted partners. developers, businesses and consumers come later, "as soon as possible", with no date.
update: the price story
a few days later, 9to5google covered a new benchmark of 206 work tasks across finance, legal, software and more.
claude opus 5.5 came out on top. gemini 4 carbon was very close behind. the gap that matters is cost: around $51 per task for opus against just over $16 for carbon.
the same report notes the ranking shuffles when some domains are taken out, so treat it as one benchmark, not the final word.
why it matters for a creator business
if you build anything heavy in house, like a reporting backend or tooling for your team, model cost per task is a real line in your budget.
a model that's close to the best at a third of the price is worth a test the day it opens up.
the daily post, by email. nothing else.
one email a day max. unsubscribe anytime.