Google Just Released Its Most Advanced Audio Model. Here Is How It Ranks
Google launched Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on September 15, calling them its most superior audio fashions but.
The two fashions speak, purpose, and deal with duties. But how do they rank with unbiased analysts? The Extended Thinking model topped Artificial Analysis’ Speech-to-Speech Index at 82.6, forward of GPT-Live-1 and Grok Voice.
Follow us on X to get the most recent information because it occurs
How the Benchmarks Rank Google’s Newest Conversational AI Model
Artificial Analysis’s Index averages speech reasoning, agentic efficiency, enviornment choice, and job success charge.
Gemini 3.8 Live Extended Thinking, examined at high reasoning effort, debuted in first place. GPT-Live-1 Astra adopted at 81.5 and Grok Voice Think Fast 2.0 High at 81.3.
The normal Gemini 3.8 Live placed fifth with a score of 76.0. Both variants beat Gemini 3.1 Flash Live High, which scored 71.5.
The agentic hole is wider. Extended Thinking reached 68.6% on the Tau Voice benchmark, in opposition to 37.7% for the earlier era.
On the speech reasoning benchmark, Extended Thinking scored 97.7%, edging Grok Voice at 97.2%. However, it trailed Qwen Audio 3.0 Realtime Plus, which scored 99.2%.
Human Testers Still Reach for the Older Gemini Model
Price is the place the space opens up. The normal mannequin runs $0.84 per hour of enter audio. That is roughly half its predecessor’s $1.75 and the Index’s most cost-effective charge.
Extended Thinking runs at $3.50 per hour. That undercuts GPT-Live-1 Sol at $4.47 and Grok Voice Think Fast 2.0 High at $4.80.
Latency fell as properly. Average time to first audio dropped to 1.18 seconds, in contrast with 2.99 seconds for the older Gemini mannequin.
Listeners, nonetheless, should not totally bought. Gemini 3.1 Flash Live nonetheless leads in choice in blind Speech Agent Arena conversations, with an Elo of 1096.
Gemini 3.8 Live sits second at 1083. The Extended Thinking variant trails at 990, regardless of finishing 89.1% of its duties.
Voice AI is now being graded on two scales that time in numerous instructions. Benchmarks reward the mannequin that causes hardest. Preference rewards the one who talks greatest. Google is presently leading both, with different fashions.
Subscribe to our YouTube channel to look at leaders and journalists present knowledgeable insights
The publish Google Just Released Its Most Advanced Audio Model. Here Is How It Ranks appeared first on BeInCrypto.
