Parameters
-
Context Length
2.1M
Modality
Multimodal
Architecture
Dense
License
Proprietary
Release Date
8 Jan 2026
Knowledge Cutoff
Oct 2025
Rank
#10
| Benchmark | Score | Rank |
|---|---|---|
Professional Knowledge | 0.90 | 🥈 2 |
Graduate-Level QA | 0.919 | 🥈 2 |
StackUnseen | 0.862 | 7 |
Data Analysis high | 0.79 | 13 |
General Text | 1486 | ⭐ 14 |
General high | 0.77 | 17 |
LiveBench Average high | 0.77 | 20 |
Mathematics high | 0.91 | 29 |
Reasoning high | 0.84 | 36 |
Coding high | 0.76 | 37 |
Agentic Coding high | 0.44 | 42 |
Web Development | 1438 | 64 |
Overall Rank
#10
Coding Rank
#20
Gemini 3 Pro Preview is Google DeepMind's high-capacity multimodal model designed for enterprise data processing and cross-modal reasoning. It delivers robust tool orchestration and code generation over an expansive 2M token context window.
Detailed architecture specifications are limited for proprietary models.
Google's latest generation multimodal models with breakthrough performance across coding, mathematics, reasoning, and language understanding. Features ultra-large context windows, native multimodal processing, and thinking modes with minimal latency overhead. Available in Pro and Flash variants optimized for different workloads, with preview versions showing state-of-the-art results on multiple benchmarks.
APX AI
Online