“The biggest model upgrade to Search to date”: Gemini 4 Hydrogen rolls out fully, 3.8 Flash retires directly
On November 12, Google launched Gemini 4 Hydrogen, making it available worldwide starting today to all Google Search users, the Gemini app, Gemini API, AI Studio, Gemini for Business, and Antigravity.
Hydrogen is the execution-focused model in the Gemini 4 family, complementing the frontier reasoning models Carbon and Argon, and directly replacing Gemini 3.8 Flash.
Compared with 3.8 Flash, it completes tasks 2.5 times faster, scores 57% on Terminal-Bench 4.1 versus just 18.2% for 3.8 Flash, and produces over 5 times fewer fabricated claims.
On the Artificial Analysis Intelligence Index, a third-party benchmark, it scores 51 at the default high setting—above Argon’s 50—with a score of 54 at the xhigh setting.
API pricing is standardized across the full 1M context window at $1.50 per million input tokens and $7.50 per million output tokens.
On the day the news broke, Google shares closed up 1.79% at $298.70.
This single upgrade moves three needles at once: speed, hallucinations, and price. Search is Google’s cash cow, and putting the model directly in front of all Search users means bringing its strongest execution capabilities to the largest audience. If the stock holds above $300 over the next month, that would be a positive sign; if Search revenue growth slows in the next earnings report, or a competitor’s new model surpasses Hydrogen’s benchmark scores, this narrative could quickly lose momentum.
Risks: Hydrogen is only an execution-focused model; the true frontier reasoning models, Carbon and Argon, have not yet been integrated into Search. Benchmark scores come from a third party, and real-world experience may fall short.
$GOOGL
#Gemini #Google
On November 12, Google launched Gemini 4 Hydrogen, making it available worldwide starting today to all Google Search users, the Gemini app, Gemini API, AI Studio, Gemini for Business, and Antigravity.
Hydrogen is the execution-focused model in the Gemini 4 family, complementing the frontier reasoning models Carbon and Argon, and directly replacing Gemini 3.8 Flash.
Compared with 3.8 Flash, it completes tasks 2.5 times faster, scores 57% on Terminal-Bench 4.1 versus just 18.2% for 3.8 Flash, and produces over 5 times fewer fabricated claims.
On the Artificial Analysis Intelligence Index, a third-party benchmark, it scores 51 at the default high setting—above Argon’s 50—with a score of 54 at the xhigh setting.
API pricing is standardized across the full 1M context window at $1.50 per million input tokens and $7.50 per million output tokens.
On the day the news broke, Google shares closed up 1.79% at $298.70.
This single upgrade moves three needles at once: speed, hallucinations, and price. Search is Google’s cash cow, and putting the model directly in front of all Search users means bringing its strongest execution capabilities to the largest audience. If the stock holds above $300 over the next month, that would be a positive sign; if Search revenue growth slows in the next earnings report, or a competitor’s new model surpasses Hydrogen’s benchmark scores, this narrative could quickly lose momentum.
Risks: Hydrogen is only an execution-focused model; the true frontier reasoning models, Carbon and Argon, have not yet been integrated into Search. Benchmark scores come from a third party, and real-world experience may fall short.
$GOOGL
#Gemini #Google