Google is making moves with its next generation of artificial intelligence, starting the early training phase for Gemini 4. This major step forward happens while the current flagship model, version 3.5 Pro, remains stuck in partner testing. The shift in focus raises questions about whether the tech giant can keep pace with rapid advancements from heavy hitters like OpenAI and Anthropic right now.
The company begins early training for its next big model
The next generation of AI is officially in the works. The search giant confirmed that pre-training for its version 4 model has started. This foundational stage involves feeding massive datasets into the system to teach it language, coding, and basic reasoning skills before more advanced fine-tuning begins.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
While the company did not provide technical details or a specific release date, the move shows a clear push toward building a more capable system. The goal is to establish a strong presence in the race toward general artificial intelligence.
The current flagship model remains delayed in partner testing phases
The push for a new version comes as the 3.5 Pro model continues to miss its launch targets. Originally announced in May and expected in June, the flagship reasoning model is still undergoing rigorous testing with select enterprise partners.
Reports indicate the delay stems from efforts to improve coding performance and ensure stability. Instead of rushing a product to market, the teams are taking extra time to validate safety protocols and iron out any issues.
New efficiency models launch to bridge the gap for developers
To keep developers engaged during the wait, the company rolled out three new Flash models. These include 3.6 Flash, 3.5 Flash-Lite, and a security focused 3.5 Flash Cyber.
These options are built for speed and token efficiency. By offering cheaper and faster tools right now, the company hopes to maintain its user base while the heavy lifting happens behind the scenes on the larger models.
The strategy of teasing a future breakthrough while delaying a current product carries real risks. Developers need reliable tools today, not just promises for tomorrow. The upcoming months will reveal if this balancing act pays off or if competitors pull further ahead in the race.
Discussion