google/gemini-3.1-pro

Google's most intelligent model, with improved reasoning and a new medium thinking level

393 runs

Readme

Gemini 3.1 Pro

Overview

Gemini 3.1 Pro is the next iteration in the Gemini 3 Pro family, delivering improved performance, behavior, and intelligence. It builds on Gemini 3 Pro’s strengths in reasoning, coding, and multimodal understanding.

What’s new in 3.1 Pro

  • Medium thinking level: A new balanced reasoning mode between low and high, giving you finer control over latency vs. reasoning depth.
  • Improved intelligence: Better performance across reasoning, coding, and complex multimodal tasks compared to Gemini 3 Pro.
  • Same great features: 1 million token context window, 64k token output, multimodal input (text, images, video, audio), and system instructions.

Thinking levels

Gemini 3.1 Pro supports three thinking levels that control how deeply the model reasons before responding:

Level Best for Latency
low Simple tasks, chat, high-throughput applications Fastest
medium Balanced reasoning for most tasks Moderate
high (default) Complex reasoning, code audits, strategic analysis Slowest

Inputs

  • prompt: Text prompt to send to the model
  • images: Up to 10 images (each up to 7MB)
  • videos: Up to 10 videos (each up to 45 minutes)
  • audio: One audio file (up to 8.4 hours)
  • system_instruction: Guide the model’s behavior
  • thinking_level: Control reasoning depth (low, medium, high)
  • temperature: Sampling temperature (0-2, default 1.0)
  • top_p: Nucleus sampling parameter (0-1, default 0.95)
  • max_output_tokens: Maximum tokens to generate (up to 65,535)

Pricing

Pricing is per token, with two tiers based on context length:

Context Input Output
≤200k tokens $2 / 1M tokens $12 / 1M tokens
>200k tokens $4 / 1M tokens $18 / 1M tokens
Model created
Model updated