Gemini 4 Argon Launches, Plus Bing's New Topics Portfolio

TL;DR
Google launched Gemini 4 Argon this week across the API, Google Antigravity, the Gemini app, and AI Mode in Search, pricing it at $2 per million input tokens and $10 per million output tokens. In an 11-hour overnight run, Argon also optimized Google's own 205,000-line ADK codebase, cutting session prep time by about 40% and speeding cold imports by 21%. Separately, Microsoft added a beta Topics Portfolio feature to Bing Webmaster Tools that groups AI Search citations into thematic clusters.
Key takeaways
- Gemini 4 Argon is live now via the API in AI Studio, Google Antigravity, the Gemini app, and AI Mode in Search.
- Argon ranks #1 on the Vals Index, a benchmark covering finance, coding, legal, and tax work.
- Argon costs $2 per million input tokens and $10 per million output tokens, with cached input tokens 95% off.
- In an 11-hour overnight run, Argon cut ADK's in-memory session prep time by about 40% and sped up cold imports by 21%.
- Bing Webmaster Tools added a beta Topics Portfolio that groups AI Search grounding queries into thematic clusters.
Google: Gemini 4 Argon Launches
Google released Gemini 4 Argon today, available via the API in AI Studio and the Gemini platform, for Google AI Ultra subscribers in Google Antigravity, and for Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search. Argon ranks #1 on the Vals Index, which measures model performance across finance, coding, legal, and tax work. It launches at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off.
- Live now via API in AI Studio and the Gemini platform
- Ranks #1 on the Vals Index
- $2 per million input tokens, $10 per million output tokens
- Cached input tokens are 95% off
Why it matters: A cheaper, benchmark-leading model inside Gemini, AI Mode in Search, and Google Antigravity gives brands another surface where an AI engine shapes how they're described, so what Argon says about a company now carries more weight.
Source: Google: Gemini 4 Argon: Built to solve your hardest problems (2026-10-13)
Google: Argon Optimizes Code Overnight
Google ran Gemini 4 Argon overnight on google/adk-python, a roughly 205,000-line Python codebase with nearly 800 source files and nearly 17,000 unit tests. In an 11-hour unattended run, Argon cut ADK's in-memory session prep time by about 40% in long sessions and sped up cold package imports by 21%.
- google/adk-python: ~205,000 lines, nearly 800 files, nearly 17,000 unit tests
- 11-hour overnight run
- Cut in-memory session prep time by about 40%
- Sped up cold package imports by 21%
Why it matters: It shows Argon can handle long, unattended agentic work on a real production codebase rather than just short demo tasks, which matters for any team weighing how much AI-visibility tooling or content work they can hand to an agent.
Source: Google: Optimizing a 200,000-line Python codebase overnight with Gemini 4 Argon (2026-10-13)
Microsoft: Bing Adds Topics Portfolio
Microsoft added a beta feature called Topics Portfolio inside the AI Performance report in Bing Webmaster Tools. It groups the grounding queries where a site appears into thematic clusters and shows Topic Size, Sites Cited, and Queries for each one, plus a Top Opportunity Queries list of strong queries where a site isn't showing up.
- New beta tab in Bing Webmaster Tools' AI Performance report
- Groups grounding queries into thematic clusters
- Shows Topic Size, Sites Cited, and Queries per topic
- Includes a Top Opportunity Queries list
Why it matters: This gives brands a direct, topic-level view of where they're already cited in Bing's AI answers and where competitors are cited instead, turning AI visibility into something that can be planned by theme rather than tracked query by query.
Source: Microsoft: Introducing Topics Portfolio (Beta) in AI Performance (2026-10-13)
What this means for your brand's AI visibility
We're watching Gemini 4 Argon closely because it now answers inside AI Mode in Search and the Gemini app, which means its read on a brand carries into the same surfaces where people check reviews and compare products. If Argon ranks #1 on a benchmark like the Vals Index, buyers researching finance, legal, or technical purchases may lean on it more, so the facts it has about a brand are worth checking now rather than after a bad answer spreads.
The Bing update matters just as much for how we plan content. A Topics Portfolio that shows which themes a site already owns, and which opportunity queries are open, gives us a concrete way to prioritize the next piece of content instead of guessing. We'd rather point a client at an open query with real citation volume than write another generic brand page. Now that Argon is live across Search and the Gemini app, it's worth learning how to track how Gemini mentions your brand. Bing's Topics Portfolio is one more reason to improve your brand's visibility in AI search engines before competitors claim the open queries. We built the Sophyx AI Visibility Tracker to surface exactly these kinds of citation gaps across platforms like Bing and Gemini. Benchmark wins like Argon's top Vals Index ranking can shift your AI share of voice overnight, so it's worth checking where you stand.
Frequently asked questions
What is Gemini 4 Argon?
Gemini 4 Argon is Google's new frontier model, now available via the API in AI Studio and the Gemini platform, through Google Antigravity, and in the Gemini app with AI Mode in Search.
How much does Gemini 4 Argon cost?
It launches at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off.
How well does Gemini 4 Argon perform on benchmarks?
Argon ranks #1 on the Vals Index, which measures model performance across finance, coding, legal, and tax work.
What is Bing's new Topics Portfolio feature?
Topics Portfolio is a beta feature in the AI Performance report inside Bing Webmaster Tools that groups the grounding queries where a site appears into thematic clusters, showing Topic Size, Sites Cited, and Queries for each one.
Video transcript
Gemini 4 Argon is Google's big news this week, a new model for coding, finance, legal, and science. Argon is live now via the API, Google Antigravity, and the Gemini app with AI Mode in Search. It ranks number one on the Vals Index, which grades models on finance, coding, legal, and tax work. Pricing starts at $2 per million input tokens and $10 per million output tokens. Overnight, Argon optimized Google's 205,000 line ADK codebase without any help. It cut session prep time by about 40% and sped up cold imports by 21%. Microsoft's Bing Webmaster Tools now has a beta feature called Topics Portfolio. It groups your AI citations into themes, showing where you're winning and where rivals show up. Follow Sophyx for your daily AI platform update, and read the full brief on our blog.
Related on Sophyx
- track how Gemini mentions your brand
- improve your brand's visibility in AI search engines
- Sophyx AI Visibility Tracker
- AI share of voice
- Yesterday's update: Google's Universal Cart and Anthropic's New AI Evaluators
Sources are linked above. Written by Reza Kashi from each company's own announcement; we update this brief every weekday morning.
Comments
Loading comments…