**Major improvements in Grok 4.1 (released November 17, 2025):**- **Significantly faster responses** with lower latency (up to 42% in some reports) and higher overall speed.- **Sharply reduced hallucinations** (3x lower rate, from ~12% to ~4%, for better factual reliability).- **Top-tier conversational quality**: #1 on LMArena Text Arena (1483 Elo for Thinking mode, 31-point lead over rivals), 65% user preference in blind tests.- **Breakthrough emotional intelligence & creativity**: Record scores on EQ-Benc ...

RT Tech Dev Notes (@techdevnotes)Grok 4.1 Thinking is so good at poems, creative writing and does not hold back at all, it almost writes anythingIt's way better than Grok 4 Thinking or Grok 4 Fast ...

RT DogeDesigner (@cb_doge)Here are the key improvements in Grok 4.1 as compared to its previous models:▸ Better user preference: In blind pairwise tests during rollout, Grok 4.1 was preferred ~64.78% of the time over the previous production model.▸ Enhanced emotional and interpersonal ability: It performs stronger on emotional-intelligence benchmarks (e.g., EQ-Bench) and is more capable at nuanced, empathetic responses.▸ Improved creative writing and style: In benchmarks for creative writing, it shows more ...

🧵 How to use Grok 4.1 like a pro (2025 edition) 🚀1/ Grok 4.1 is now FREE for everyone – no paywall!Just go to https://t.co/KaH5w8Ke4N, https://t.co/SNtzbLKMKG, or the Grok/X apps and select “Grok 4.1” in the model picker (or leave it on Auto).Thinking mode = deep reasoning. Fast mode = instant replies. You get both.2/ Be brutally specific with promptsBad: “Tell me about quantum computing”Good: “Act as a PhD physicist who taught Feynman. Explain quantum entanglement in 3 levels: (1) 5-year-old, (2) college f ...

RT Mark Kretschmann (@mark_k)Grok 4.1 first impressions:* Very different writing style, much more personal.* Much improved image understanding, sees even small details.* Comes in two flavors, normal and "thinking".* Quite fast, even the thinking version. ...

Core Insights - The launch of Grok 4.1 by xAI, led by Elon Musk, has positioned it as the new leader in AI models, surpassing competitors like Gemini 2.5 Pro in performance and capabilities [1][3][9] Performance Metrics - Grok 4.1 Thinking achieved a score of 1483 Elo, making it the top model globally, outperforming Gemini 2.5 Pro by 31 points [3][13] - The non-reasoning version of Grok 4.1 scored 1465 Elo, placing it second overall [3][14] - Grok 4.1 has shown a significant improvement in writing capabilities, with a 600-point increase in Elo compared to its predecessor [8][60] Emotional Intelligence - Grok 4.1 has demonstrated enhanced emotional intelligence, achieving a score of 1586 Elo on the EQ-Bench3, which measures emotional understanding and interpersonal skills [15] - The model's ability to engage in natural and empathetic conversations has been highlighted, showcasing its improved interaction capabilities [7][11] Technological Advancements - The xAI team expanded the reinforcement learning (RL) scale during the post-training phase, contributing to Grok 4.1's rapid evolution [8][11] - The hallucination rate of Grok 4.1 has decreased by three times compared to the previous model, indicating improved factual accuracy [8][60] User Preference - In blind tests, users preferred Grok 4.1 over its predecessor in 64.78% of cases, reflecting its enhanced performance and user experience [11][60]

Core Insights - Elon Musk's AI company xAI has launched its latest large language model, Grok 4.1, which is now available to all users on grok.com and mobile applications [1][2] Performance and Capabilities - Grok 4.1 aims to enhance usability in real-world scenarios, showing significant improvements in creativity, emotional understanding, and collaborative interaction [2] - The model achieved top-tier performance in the LMArena text capability leaderboard, with its deep-thinking version (quasarflux) scoring 1483 Elo, leading the second-place model by 31 points [4] - The "instant response" version of Grok 4.1 scored 1465 Elo, outperforming all other models in "full reasoning" mode, marking a substantial leap from the previous Grok 4, which ranked 33rd [4] Emotional and Creative Intelligence - Grok 4.1 demonstrated significant advancements in "soft skills," excelling in the EQ-Bench3 benchmark for emotional intelligence and the Creative Writing v3 test for creative capabilities [5][6] - In the EQ-Bench3 test, Grok 4.1's reasoning and non-reasoning modes secured the top two positions, while in creative writing, it ranked second and third, just behind the earlier GPT-5.1 model [6] Reliability and Accuracy - The model has improved its ability to handle complex logical reasoning and better understand emotional prompts, enhancing its human-like interaction [9] - A key improvement is the significant reduction in the model's "hallucination" rate, which refers to factual inaccuracies, particularly in fast-response models equipped with search tools [10] - The hallucination rate has been notably decreased, providing users with more reliable and accurate information [10]

梦晨 发自 凹非寺 量子位 | 公众号 QbitAI 刚刚，马斯克发布Grok 4.1，同时霸榜大模型竞技场的第一和第二。 怎么做到的？ Grok 4.1思考模式 以1483的Elo分数稳居榜首，领先非xAI模型中的最高分整整31分。 Grok 4.1非思考模式 以1465分拿下第二名，超越了公开排行榜上所有其他模型的完整推理模式。 | Rank 14 | Rank Spread O (Upper-Lower) | Model 14 | Score ↓ | 95% Cl (±) 11 | Votes 11 | Organization 1J | License 11 | | --- | --- | --- | --- | --- | --- | --- | --- | | 1 | 1 4-12 | X grok-4.1-thinking | 1483 O Preliminary | ±11 | 3,298 | ×AI | Proprietary | | 2 | 1 < > 4 | XI grok-4.1 | 1465 O Preliminary | ±11 | 3,413 | ×AI | Proprietar ...

🚨 Grok 4.1 just dropped! 🚀What's new:- Way smarter in creative writing, emotional chats, and collab vibes 😎- Picks up on subtle nuances, stays super coherent & fun- 3x fewer hallucinations + noticeably faster responses- Tops LMSYS Arena: #1 with 1483 Elo (Thinking mode) 🔥Available NOW to everyone on https://t.co/KaH5w8Ke4N, X, iOS/Android apps – rolling out in Auto mode!Upgrade your convos today. Who's trying it first? 👀 ...

RT X Freeze (@XFreeze)xAI just quietly released a SOTA model that’s insanely powerfulDominates the LM Arena Text leaderboard, securing both #1 and #2 spots - the highest scores everHallucination rates are reduced by nearly 70% compared to the previous version, setting a new standard for reliabilityGrok 4.1 now finds images, includes them in responses, and can search and play YouTube videos right in the chatRanks at the top on the Emotional Intelligence Benchmark - understands human emotions much better and ...