Product Launchllmreasoninggooglebenchmarks
Google Releases Gemini 3 Deep Think
9.0
Google unveiled Gemini 3 Deep Think on Dec 5, 2025, an advanced reasoning mode for Gemini 3 Pro available to Ultra subscribers. It claims parallel multi-hypothesis reasoning and reports 41.0% on Humanity’s Last Exam and 45.1% on ARC-AGI-2 with code execution. The feature aims to improve multi-step problem solving for researchers, developers, and students via the Gemini app.
Key Points
- 1Introduces parallel hypothesis exploration in Gemini 3 Deep Think, enabling simultaneous multi-path reasoning.
- 2Demonstrates record benchmarks: 41.0% on Humanity’s Last Exam and 45.1% on ARC-AGI-2 with code.
- 3Offers Ultra subscribers in Gemini app a practical co-pilot for researchers, developers, and students.
Scoring Rationale
Strong official benchmarks and feature release, limited by gated Ultra-only access and lack of independent external validation.
Sources
Public references used for this report.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
Active Search Campaigns by BudgetEasyHigh CPC Clicks & Poor Landing PagesMediumCampaign ROAS by Attribution ModelHard
250 free problems · No credit card
See all Ad Tech problems
