💻 Technology May 6, 2026 · Ryan Whitwam

Google's Gemma 4 AI models get 3x speed boost by predicting future tokens

#AI#Google#Gemma
Ars Technica
Technology and science news
View Channel →
Google's Gemma 4 AI models get 3x speed boost by predicting future tokens
Source ↗ 👁 8 💬 0
Google launched its Gemma 4 open models this spring, promising a new level of power and performance for local AI. Google's take on edge AI could be getting even faster already with the release of Multi-Token Prediction (MTP) drafters for Gemma. Google says these experimental models leverage a form of speculative decoding to take a guess at future tokens, which can speed up generation compared to the way models generate tokens on their own.
The latest Gemma models are built on the same underlying

Comments (0)

Sign in to join the discussion

More Like This

📰
Intel Razor Lake 12-core P-Core CPU leaks with bLLC cache tech
OC3D · 11m ago
Raj Shamani Becomes World’s Most-Subscribed YouTube Podcaster
Inc42 Media · 47m ago
Analyst claims 38 million Xbox Series X|S consoles have been sold so far
KitGuru · 55m ago
Auxilion to create 50 new jobs in Dublin, Belfast and Sheffield
Silicon Republic · 1h ago
iPhone Duo 2’s biggest upgrade is racing against the clock
Cult of Mac · 1h ago
📰
Jensen Huang considers Nvidia CEO comeback – Report claims
OC3D · 1h ago