Google DeepMind is targeting July 17 for the general availability of Gemini 3.5 Pro, a model the company chose to rebuild entirely from scratch after early enterprise testing revealed performance gaps in mathematical reasoning, SVG generation, and image quality.
The original Gemini 3.5 Pro architecture, an evolution of the 2.5 Pro model, could not close these gaps through fine-tuning alone. Rebuilding from scratch is an unusual call that extends the timeline but avoids carrying forward architectural limitations. Internally, Google has reportedly framed the decision as unwillingness to ship a model that would fall short against Fable 5 and GPT-5.6.
Expected specifications from enterprise preview channels include a 2-million-token context window, double the 1 million cap on Gemini 2.5 Pro, a Deep Think reasoning mode for extended problem solving, and autonomous workflow capabilities. Pricing is expected at $15 per million input tokens and $60 per million output tokens, positioning it as a premium-tier offering.
The launch arrives in a compressed ten-day window alongside DeepSeek V4’s stable release on July 24 and ongoing limited availability of GPT-5.6 Sol.
