< Back to all clusters
[TECHNOLOGY] · United States · 16 sources

started · updated

Google ramps up compute for Gemini 4 AI model launch

Google CEO Sundar Pichai said the company is allocating large amounts of Tensor Processing Unit (TPU) compute to develop Gemini 4, calling it the most ambitious pre‑training run to date. The new model will use a larger base model to stay competitive at the AI frontier and is expected to be released in November or December 2026. Development of Gemini 4 runs in parallel with the delayed Gemini 3.5 Pro, which remains in partner testing after missing its original June launch window. Google has also introduced three new Flash models—Gemini 3.6 Flash, 3.5 Flash‑Lite and 3.5 Flash‑Cyber—designed to improve coding performance, lower token usage and handle security‑focused tasks, with updates planned on an almost monthly cadence. Pichai reiterated confidence that Google will remain at the leading edge of artificial‑intelligence research.

Entities

Alphabet Inc. · Gemini 3.5 Pro · Gemini 4 · Google · Google LLC · Sundar Pichai · Tensor Processing Units (TPU)

Claims

What the coverage asserts, and how many sources carry each claim.

Sources

about 2 months ago
about 2 months ago
about 2 months ago