Google plans to release a new Gemini 3.5 Pro model featuring a record-breaking 2 million token context window and a specialized "Deep Think" mode for solving complex logical tasks. The release is expected in July 2026, and advanced features will be available to users on the premium Ultra tier.
What Happened
Google has moved the expected release of Gemini 3.5 Pro from June to July 2026. The new model will offer an expanded context window of 2 million tokens and a Deep Think mode focused on deep logical reasoning. Access to the deep thinking mode will be provided through the Ultra tier, costing $250 per month.
Context
Current market solutions, such as Claude Opus 4.8, offer a context window of 1 million tokens. Gemini 3.5 Pro effectively doubles this figure, setting a new technical standard for frontier models.
Why It Matters for the Industry
For the industry, this signifies a shift toward Native Long-Context approaches, which may reduce the need for complex RAG pipelines when processing medium volumes of data. It also sets a new benchmark in the race for working memory capacity, forcing competitors like Anthropic and OpenAI to rethink their architectural solutions for creating autonomous AI agents.
Why It Matters for Users
Users will gain the ability to process massive amounts of information—entire books, giant code repositories, or hour-long videos—in their entirety without losing context. This will significantly increase the accuracy of data analysis and transcription, making work on large projects deeper and more seamless.
What Is Not Yet Known / Limitations
At this time, there is no precise data regarding latency or API usage costs. There are concerns regarding the high total cost of ownership for the enterprise segment and uncertainty regarding performance when working with the maximum context window.
Sources
Author
Look at AI, Editorial Team