Google DeepMind Releases Gemini 3.1 Pro: 2x Reasoning Boost and Record Benchmark Scores

Original: Google DeepMind Releases Gemini 3.1 Pro: 2x Reasoning Boost and Record Benchmark Scores View original →

Read in other languages: 한국어日本語
LLM Feb 22, 2026 By Insights AI (Twitter) 1 min read 8 views Source

Overview

Google DeepMind released Gemini 3.1 Pro on February 19, 2026, delivering over 2x reasoning improvement compared to Gemini 3 Pro. The model achieves 77.1% on ARC-AGI-2 (up from 31.1%), 80.6% on SWE-bench Verified, and tops 12 of 18 tracked benchmarks — all at the same price point as its predecessor.

Benchmark Performance

  • ARC-AGI-2: 77.1% (up from 31.1%)
  • SWE-bench Verified: 80.6%
  • GPQA Diamond: 94.3%
  • LiveCodeBench Pro Elo: 2887
  • Humanity's Last Exam: 44.4%
  • #1 on 12 of 18 tracked benchmarks

Key Features

  • 1M token context window: Supports text, images, audio, and video
  • Three thinking levels: Low, Medium, High — tune latency vs. reasoning depth
  • 64K output capacity: Suited for complex, long-form tasks
  • Multimodal: Processes text, audio, images, video, and entire code repositories

Pricing and Access

Gemini 3.1 Pro maintains the same pricing as Gemini 3 Pro at $2 per million input tokens and $12 per million output tokens. Available through the Gemini API, Vertex AI, the Gemini app, and NotebookLM.

Source: Google DeepMind (@GoogleDeepMind) on X

Share:

Related Articles

Comments (0)

No comments yet. Be the first to comment!

Leave a Comment

© 2026 Insights. All rights reserved.