Gemini 3.5 Pro to ship with 2M-token context and a Deep Think layer
The rebuilt model targets mathematical reasoning, SVG scene generation and image quality, alongside a doubled context window and autonomous workflow capabilities.
MVMara Vidović
Published 7 Jul 2026, 15:02Updated 28 Jul 2026, 10:161 min read
Alongside confirming the delay to July 17, Google has detailed what the rebuilt Gemini 3.5 Pro is actually targeting: a 2 million-token context window and a new Deep Think reasoning layer for complex, multi-step problems.
The headline specs
- ▸2 million-token context window, doubling the prior generation
- ▸A dedicated Deep Think Reasoning Layer for complex problem-solving
- ▸Targeted gains in mathematical reasoning, SVG scene generation and image quality
- ▸Autonomous workflow capabilities aimed at longer-running agentic tasks
Reading between the specs
Doubling context is the easy headline number; pairing it with a dedicated reasoning layer suggests Google is trying to solve the harder problem of using that context well in long agentic runs, not just accepting longer inputs.
Source: Google DeepMind
