Skip to content
Google

Gemini 3.1 Pro

Frontier

GoogleReleased on 2026-02-19

Google's most advanced Pro-tier model with 1M context, dynamic thinking, and the highest ARC-AGI-2 score (77.1%) among all models. Excels at multimodal reasoning across text, images, audio, and video. Best price-to-performance ratio among frontier models.

86
Overall Score

Voice of the community

sample 1,450

Gemini has turned into what seems to be a retarded drunk toddler compared to Claude or GPT... 95% of the time, it feels like the context Window is just as absent as its thinking process.

r/GeminiAI2026-03-24

I am 99% sure 3.1 PRO has been nerfed as I tried many one shot apps using 3.1 PRO with different prompts... now 3.1 PRO does not generate outputs comparable to 3.0

r/Bard2026-03-26

For me Gemini 3.1 pro consistently beats others on normal conversations and tasks. So much so that I cancelled my OpenAI subscription.

r/LocalLLaMA2026-03-25

Core Specs

1000K
Context Window
66K
Max Output
ReasoningOpen Sourcetextimageaudiovideo

Pros & Cons

Sentiment42% +24% ·34% −

Pros

  • +Cheapest frontier model ($2/$12)
  • +Highest ARC-AGI-2 score (77.1%)
  • +Native multimodal (text/image/audio/video)
  • +1M context window
  • +Batch pricing 50% off ($1/$6)
  • +Best for frontend/web design

Cons

  • Weaker at agentic tasks
  • Context window issues reported
  • Sometimes skips lines in code edits
  • CLI/agent harness quality inconsistent

Pricing

Input (per 1M tokens)$2.00
Output (per 1M tokens)$12.00
Subscription$20/month
Updated on 2026-03-28

Get Started with Gemini

2Sign in with Google account
3Upgrade to Advanced ($20/mo) for Pro

💡 Gemini Advanced is included in Google One AI Premium.

Benchmarks

arcAgi277.1%
mmlu91.8%
sweBench76.2%
gpqaDiamond91.9%

Reliability

SLA99.5%
Incidents (30d)11
Last Incident2026-05-05
⚠️ 4/28 10:00 UTC 检查:4/27 06:34 UTC 起 minor warning 仍未完全解除但已显著降温 (~27.5h+),24h 用户报告 25→9,Workspace 03:32 UTC 复绿,AI Studio / Gemini API 仍标 'partial outage' (StatusGator-only,官方状态页未公告);4/28 01:00-01:13 UTC 另有 13min 独立 minor warning (已 resolved);4/26 23:29-01:24 UTC ~1h55min;4/25 22:12-04:18 UTC 6h+ Gemini API/AI Studio 全组件警告级降级;4/20 OpenAI outage 期间溢出流量短暂错误率升高。最近 24h 内累计 4 轮 warning,趋势收敛中;warning 级且非完全不可用,仍未达 Telegram 通知阈值。SWE-bench 表现强劲,用户反馈 coding 干净但 agentic 任务较弱。
View Status Page →