Gemini 3.1 Pro (deepmind.google)

Article summary

Gemini 3.1 Pro is a next-generation AI model that can comprehend vast datasets and challenging problems from multimodal information sources. The model is based on Gemini 3 Pro and has been evaluated across various benchmarks, showing significant improvements in reasoning and multimodal capabilities. Gemini 3.1 Pro is available through various channels, including the Gemini App, Google Cloud, and Google AI Studio. The model's safety and ethics have been evaluated, and it has been found to outperform its predecessor in safety and tone.

Main themes

• AI model capabilities
• Multimodal reasoning
• Safety and ethics
• Model evaluation
• Competition and innovation
• User experience and accessibility

What commenters say

• The new Gemini 3.1 Pro model has shown impressive results in generating complex images, but may not always produce photorealistic output.
• Some users have reported inconsistent results when using the model, with varying levels of success in generating images.
• Gemini 3.1 Pro has improved long context capabilities, making it suitable for tasks such as codebase migration.
• The model's ability to follow complex multi-step instructions is still limited, and may take shortcuts when faced with multiple constraints.
• Competition between AI models is driving innovation, with Gemini 3.1 Pro being closely comparable to other models such as GPT-5 and Claude Opus 4.
• Some users are experiencing issues with accessing the latest model through the Gemini CLI, despite having a subscription.
• The Gemini 3.1 Pro model has been found to have improved safety and tone compared to its predecessor, but may still have limitations and biases.