• New Gemini 3.1 Pro crushes previous benchmarks, outperforms GPT 5.2 reasoning The latest Gemini update sharpens coding support and nearly doubles performance in agentic workflow tests, signaling Google’s focus on practical AI deployment. • Google has rolled out Gemini 3.1 Pro, the latest update to its flagship AI model, just months after releasing Gemini 3 in November. • The new version enters preview today for developers, enterprises, and consumers, with Google promising stronger reasoning, better coding performance, and improved handling of long documents. • The company says Gemini 3.1 Pro powers the “core intelligence” behind recent upgrades to its Deep Think tool. • While benchmark gains appear modest in some areas, Google claims the update delivers more consistent and reliable performance in real-world tasks. • Stronger reasoning benchmarks Google highlighted performance gains across several industry tests.
Article Summaries:
- Google has launched Gemini 3.1 Pro, its latest flagship AI model, offering stronger reasoning, improved coding support and a larger context window (up to 1 million input tokens and 64,000 output tokens). In benchmark tests the update nearly doubles performance on ARC‑AGI‑2 (77.1 % vs 31.1 %) and scores 44.4 % on Humanity’s Last Exam, surpassing OpenAI’s GPT 5.2 (34.5 %). While it trails Claude Opus 4.6 on the Arena leaderboard, Gemini 3.1 Pro shows significant gains in agentic workflow tests (APEX‑Agents). Pricing remains unchanged, and the model is being rolled out across Google’s developer, enterprise and consumer platforms.
Sources: