Gloria daily close for session Oct 27, 2026. Market cap and trailing P/E use this close with the SEC share total and trailing diluted earnings, each dated next to its figure. Annual revenue, revenue growth, net income and margin show their own fiscal period and link the 10-K. Current context may postdate this article.
Alphabet Inc. and its peers
Token vs underlying stock
Current comparison for GOOGL. Historical observations may postdate this article.
GOOGLx token unit vs GOOGL share · USD
$321.43Oct 27 UTC
-15.0% over displayed period
● GOOGLx token┄ GOOGL stock: $320.58
Aug 4$316.80 to $378.19Oct 27
Alphabet CEO Sundar Pichai says Google’s new Carbon model is the strongest AI system the company has built. The full name is Gemini 4 Carbon, and Pichai introduced it on October 28, 2026.
That would be a bigger flex if Google hadn’t shipped its previous flagship, Gemini 4 Argon, only weeks earlier.
The numbers behind the claim
The headline figure comes from the v5.0 Artificial Analysis Intelligence Index. Gemini 4 Carbon scored 59, which puts it in first place on that ranking.
The runners-up are close behind. Anthropic’s Claude Opus 5.5 scored 58, and OpenAI’s GPT-6 Astra landed at 57.
The more interesting story sits in cybersecurity. Carbon posted an 88.2% score on SEC-bench Pro, a benchmark focused on finding vulnerabilities in software.
The model also scored 54.8% on ExploitGym, which tests how well a model can actually exploit weaknesses it finds. On CWE-bench, which measures the ability to patch those weaknesses, it reached 71.0%.
Bigger gains over Argon
Google’s internal testing shows Carbon generates secure code at more than double the rate of Gemini 4 Argon.
There’s also a reported 10x improvement in producing high-quality patches in Chrome.
A gated rollout through Fairwind
Google is rolling the model out through its Fairwind Program, starting with a select group of trusted cybersecurity professionals. Broader access is anticipated soon, according to the plans described in the rollout. No firm public date accompanies that expectation.
The model is also undergoing US government pre-release testing.
What it costs
Pricing for now is identical to Argon. Introductory rates are $2 per million input tokens and $10 per million output tokens.
That introductory deal won’t last. Full pricing is set to eventually double, though the timing of that change hasn’t been specified.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.