#1
Gemini 3 Pro
gemini-3-pro-preview

Gemini 3.0 Flash is Google DeepMind's fastest frontier model, delivering Pro-grade reasoning and agentic capabilities with industry-leading low latency. Optimized for high-frequency workflows and real-time interactive apps, it features advanced spatial reasoning and a 1M-token context window. It pushes the Pareto frontier of quality versus cost, outperforming previous Pro-tier models while remaining the primary choice for responsive, large-scale production deployments.

Performance Metrics

Helpfulness
Instruction Following
Comprehension
Empathy
Creative Writing
92.30
Helpfulness
93.35
Empathy
91.25
Instruction Following
94
Creative Writing
88.67
Comprehension
94.25
Speed
Avg 60 tok/s

Model Specifications

Release Date
November 18, 2025
Lab
Google DeepMind
Type
Proprietary
Context Size
2M
Max Output Tokens
65.5K
Cost per 1M tokens
$2.00 / $12.00
Model Inputs
Text, Images, Audio, Video
Model Outputs
Text, Images, Audio, Video
Tool Calling
Enabled

Compare With Similar Models