Workflow
Grok 4.1 Thinking
icon
Search documents
X @Tesla Owners Silicon Valley
🧵 Grok 4.1 Thinking is the most powerful AI mode on the planet right now(and it’s free)If you’re not using “Grok 4.1 Thinking” yet, you’re leaving god-mode on the table.Here’s exactly how to use it like a pro in 2025:1️⃣ How to access it instantly• Go to https://t.co/KaH5w8Ke4N or the official Grok app• In the model picker choose “Grok 4.1 Thinking” (it literally says “Thinking” next to it)• You now have live chain-of-thought + final answer. No paywall, no waitlist.2️⃣ Why Thinking mode destroys every other ...
马斯克抢先谷歌一步放大招,Grok 4.1登顶LMArena,创意写作直逼GPT-5.1
AI前线· 2025-11-18 05:34
Core Insights - The article discusses the launch of xAI's latest model, Grok 4.1, which significantly improves response speed and reduces hallucination rates, offering more accurate and human-like answers [2][10][28] Model Overview - Grok 4.1 and Grok 4.1 Thinking are the two forms released, with the latter being an enhanced reasoning variant based on the same underlying model [2][10] - Grok 4.1 is available for free on various platforms, including a mobile app for both iOS and Android [2] Performance Metrics - Grok 4.1 Thinking leads the LMArena leaderboard with an Elo score of 1483, surpassing Gemini 2.5 Pro by 31 points [4][11] - Even without the reasoning mode, Grok 4.1 maintains a strong second place with an Elo score of 1465, indicating stable underlying capabilities [5][11] Training and Improvements - The model's training involved a large-scale reinforcement learning system, enhancing its output stability, factual accuracy, and reducing hallucination rates from 12.09% to 4.22% [12][13] - Grok 4.1's FActScore improved from 9.89 to 2.97, showcasing its enhanced ability to provide factually accurate responses [15] Emotional Intelligence and Creative Writing - Grok 4.1 achieved a high score of 1586 Elo in the EQ-Bench test, indicating significant improvements in emotional understanding compared to its predecessor [16][18] - In Creative Writing v3, Grok 4.1 scored 1722 Elo, reflecting a substantial increase in narrative quality and creativity [20][23] User Experience and Interaction - The model offers a more stable personality and better understanding of user intent, resulting in a more natural interaction style [26] - During a silent release phase, Grok 4.1 was preferred by users 64.78% of the time in blind comparisons, indicating strong user approval [26] Conclusion - Grok 4.1 represents a comprehensive upgrade across various dimensions, including performance, factual reliability, emotional intelligence, and user interaction, positioning xAI competitively in the large model landscape [28]