Product Launchmultimodalllmagentsbenchmark
Google Unveils Gemini 3 Deep Think Reasoning Mode
10.0
Google today announced Gemini 3 Deep Think, a new multimodal reasoning mode for the Gemini app, available to Ultra subscribers. Google says Deep Think explores parallel hypotheses and achieves 41.0% accuracy on Humanity’s Last Exam and 45.1% on ARC‑AGI‑2 (with code execution), outperforming prior public models and enabling complex planning, coding, and scientific workflows.
Key Points
- 1Launches Gemini 3 Deep Think, a multimodal reasoning mode for Ultra subscribers, handling text, images, code
- 2Achieves 41.0% on Humanity's Last Exam and 45.1% on ARC‑AGI‑2, surpassing previous public LLMs
- 3Enables complex planning, multistep problem solving, and advanced coding workflows for practitioners
Scoring Rationale
Official Google release and strong benchmark gains justify high impact, limited by gated Ultra availability and experimental status.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems
