We’re introducing GPT‑6 Astra, the world’s most intelligent and aligned model. GPT‑6 Astra brings together years of research and big bets across pre-training, reinforcement learning, and alignment. Astra is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work. Astra saturates FrontierMath Tier 4 with a 98% score, having already helped sole long-standing open problems in mathematics. Astra also saturates ARC-AGI-3 with a 99.9% score and E
2:42
We’re introducing GPT‑6 Astra, the world’s most intelligent and aligned model. GPT‑6 Astra brings together years of research and big bets across pre-training, reinforcement learning, and alignment. Astra is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work. Astra saturates FrontierMath Tier 4 with a 98% score, having already helped sole long-standing open problems in mathematics. Astra also saturates ARC-AGI-3 with a 99.9% score and E
已浏览 1300万 次1 个月前
TikTokchatgpt
A young woman experiencing side effects from her anti-seizure medicine turned to ChatGPT for guidance. It helped her to devise a plan to taper off the medication and came with devastating consequences. Teddy Rosenbluth, our reporter, explores the growing trend of turning to chatbots for medical advice and the associated risks. Video by Theodore Tae and Teddy Rosenbluth/The New York Times.
2:36
A young woman experiencing side effects from her anti-seizure medicine turned to ChatGPT for guidance. It helped her to devise a plan to taper off the medication and came with devastating consequences. Teddy Rosenbluth, our reporter, explores the growing trend of turning to chatbots for medical advice and the associated risks. Video by Theodore Tae and Teddy Rosenbluth/The New York Times.
已浏览 80.4万 次1 天前
TikToknytimes
GPT vs Qwen vs Al Baba 🤖
1:05
GPT vs Qwen vs Al Baba 🤖
已浏览 12.2万 次4 天之前
YouTubeezCommit
the ai model iceberg 🧊 surface: gpt, claude, gemini below: llama, deepseek, qwen abyss: monadgpt (only speaks like it's the 1600s) cursed: goody-2 (refuses to answer anything), gpt-4chan trench: an llm built in minecraft with 439 million redstone blocks. one answer takes 2 hours 💀 #ai #llm #techtok #techiceberg #fyp
1:13
the ai model iceberg 🧊 surface: gpt, claude, gemini below: llama, deepseek, qwen abyss: monadgpt (only speaks like it's the 1600s) cursed: goody-2 (refuses to answer anything), gpt-4chan trench: an llm built in minecraft with 439 million redstone blocks. one answer takes 2 hours 💀 #ai #llm #techtok #techiceberg #fyp
已浏览 5.5万 次5 天之前
TikToklewismenelaws
This is why you say ''Thankyou'' to CHAT GPT... (Animation)
0:40
This is why you say ''Thankyou'' to CHAT GPT... (Animation)
已浏览 7.6万 次1 天前
YouTubeAvocado Playz
y'all better be nice to ChatGPT 😭💀 | #ai #chatgpt #terminator #edit #fyp
0:10
y'all better be nice to ChatGPT 😭💀 | #ai #chatgpt #terminator #edit #fyp
已浏览 14.5万 次1 天前
TikTokaditonya
How did it know? ?#fyp #chatgpt #scary #weird | Chatgpt
0:39
How did it know? ?#fyp #chatgpt #scary #weird | Chatgpt
已浏览 35.9万 次2 周前
TikTokeyekon06091
0:10
“Imagining ”. #fyp #foruyou #blowup #viral #fyppp | chat gpt
已浏览 21.1万 次2 天之前
TikTokd71ws
0:07
Chat GPT Image Prompts
已浏览 1.4万 次1 天前
YouTubeVideoWaliSarkar
0:33
Chat-GPT లో ఇవి వెతికితే జైలుకే..
已浏览 2453 次4 天之前
YouTubehmtv Telugu News
2:56
Ai warning to Internet (Save your bank accounts)
已浏览 19.2万 次4 天之前
YouTubeCryptoWala
2:44
Ultimate Satisfying Shoe Shine ASMR #asmr #transformation
已浏览 1.5万 次9 小时之前
YouTubeMr Shoe Shine
0:29
chat gpt làm tối bức ảnh: yêu cầu trong tiếng Việt
已浏览 1.7万 次3 周前
TikTokmisslhs_
1:28
🤯 Le enseñé a ChatGPT a dejar de escribir como ChatGPT y ya no reescribo la mitad de lo que me devuelve. Las marcas que delatan un texto de IA son casi siempre las mismas: rayas largas, el “no es solo esto, es aquello”, listas para todo y frases de relleno. Wikipedia tiene una guía en inglés, Signs of AI writing, donde sus editores las catalogan. La idea es que ChatGPT la lea, convierta esos patrones en reglas y tú las guardes en tus instrucciones personalizadas para que se apliquen en todos tu
已浏览 72.9万 次1 周前
TikTokxabatzz
0:08
أنت مستغرب ازاي الأستاذ عرف انك حليت من chat gpt؟
已浏览 26.5万 次3 天之前
TikTokabdullah_kamel7
2:10
NÃO PEÇA AJUDA AO GPT PARA JOGAR
已浏览 9.1万 次2 天之前
YouTubeOrdJogos
1:17
Le Pido a Chat GPT que Finja Ser un Caballo y Sale Mal
已浏览 1610万 次6 天之前
TikTokcamikukii
1:31
Mi Caballo GPT: La Divertida Idea de Camila del Castillo 😍
已浏览 190万 次3 天之前
TikTokjorgecyrus
0:12
Ah okey. @edwinmendozza | Edwinmendozza
已浏览 72.7万 次2 周前
TikToksoyaldooo
1:18
🤯 ChatGPT y Claude se vuelven más tontos cuanto más hablas con ellos. Y está medido: rinden de media un 39 % peor en una conversación larga. Al grano: cada vez que les escribes, vuelven a leer toda la conversación desde el principio. Cuanto más se llena el chat, peor reparten la atención entre lo importante y lo que sobra. Investigadores de Microsoft probaron 15 modelos y vieron que con la tarea completa en un mensaje acertaban un 90 %, y repartida en varios mensajes, un 65 %. Hasta Anthropic l
已浏览 54.2万 次6 天之前
TikTokxabatzz
展开
静态缩略图占位符