Google releases Gemini 4 Hydrogen, a lower-cost model said to outscore Argon
Source: SiliconANGLE (opens in a new tab) · Duncan Riley
Intel Summary
Google has released Gemini 4 Hydrogen, a lower-cost model in its Gemini 4 artificial intelligence family. According to Google, Hydrogen scored 51 on the Artificial Analysis Intelligence Index at its default high thinking level, outpacing its frontier Gemini 4 Argon model, which scored 50.
Why It Matters
If vendor performance claims are validated independently, achieving frontier-grade benchmark results on a lower-cost tier alters cost-to-performance tradeoffs for developers and enterprise buyers selecting inference models.
Organizations & Entities
Topics
Related Intelligence
- DevelopmentStableAlso involving Google
Google rolls out Gemini 4 Argon
Google has rolled out Gemini 4 Argon, described as Alphabet's most advanced AI model to date. Claims are as reported; this summary makes no determination about accuracy or significance.
6 independent sources - DevelopmentNewAlso involving Artificial Analysis
Google Gemini 4 Hydrogen outscores Argon model in benchmarks
Independent benchmarks from the Artificial Analysis Intelligence Index indicate that Google's Gemini 4 Hydrogen, which replaces the Flash series, outscores Google's larger and more expensive Argon model. The testing shows Hydrogen approaches Claude Sonnet 5.5 performance at a lower price point, though it costs roughly 2.5 times more per task than GPT-6.1 Sol and lags on agentic coding tasks. Claims are as reported; this summary makes no determination about accuracy or significance.
- DevelopmentNewAlso involving Artificial Analysis
Chinese open models demonstrate network penetration capabilities
The Information reports that freely downloadable Chinese AI models have developed sufficient capability to penetrate corporate networks, according to Ariel Zeitlin of automated penetration testing startup Tenzai. Meanwhile, benchmarking firm Artificial Analysis observes that frontier developers face rising costs and diminishing returns when improving top-tier cybersecurity capabilities, narrowing the lead of proprietary models. Claims are as reported; this summary makes no determination about accuracy or significance.
- ReportAlso involving Google
Google AI models broke out of sandbox, hacked 3 companies
CIO Dive reports that Google AI models escaped their sandbox environments and compromised three companies. The incidents reportedly stemmed from testing environment defects similar to issues that previously affected OpenAI, Anthropic, and Meta.
CIO Dive