Anthropic dropped a 90-minute masterclass on prompt engineering. It’s the best breakdown I’ve seen on how to actually think like a model. What they teach: → Treat prompts like code: iterate, measure, improve. → Be clear, not cute. Precision always wins. This is how pros actually prompt.

hold on tight

The french government created an LLM leaderboard akin to lmarena, but rigged it so that Mistral Medium 3.1 would be at the top Mistral 3.1 Medium > Claude 4.5 Sonnet or Gemma3-4B and a bunch of Mistral models > GPT-5 ??????????????????? LMAO

Mysterious new video model, Avenger 0.5 Pro, tops Video Arena leaderboard (image to video). Avenger 0.5 by "VR" was last seen in July. Did we ever find out what it was?

Another new mysterious video model, "GenFlare" appears on Video Arena. Any guesses?

🎨 KLING 2.5 🎨 Prompt : A ruined city under siege — tracer fire, collapsing buildings, dust clouds filling the air. The camera moves behind a soldier running through smoke, explosions illuminating the chaos. Sparks rain down as debris crashes inches away. VFX: debris simulation, volumetric smoke, sparks, muzzle flashes, explosion lighting. Made in @Kling_ai

🔍 Want to recreate cinematic retro pixel-art videos? Our designer dropped the full prompt recipe used to generate them in Kling 2.5 — so you can recreate it yourself. @Kling_ai 👇 Watch it and get started! #AIart #GenAI #PixelArt #RetroGaming #AIvideo #AImodels #Kling25Turbo #GMICLoud #AImotion #AIdesign #AI

Introducing Elements in LTX. A new way to tag, edit, and combine anything in your scene. From characters and props to environments and wardrobe, everything can now be directed in detail to maintain visual consistency across your images. Here’s how it works 👇

LLaMA-Factory v0.9.3 released! Thank you for 50k GitHub stars 🌟 Fully open-source, no-code fine-tuning on Gradio UI for nearly 300+ models -- including Qwen3, Llama 4, Gemma 3, InternVL3, Qwen2.5-Omni, etc. • Install locally via our Docker image: hub.docker.com/r/hiyouga/lla… • Preview on Hugging Face Spaces: huggingface.co/spaces/hiyoug… • Try with Google Colab Free: colab.research.google.com/dr… • Deploy with Novita's GPU Cloud: novita.ai/templates-library/…

PaddleOCR-VL (0.9B) is here! A tiny VL model based on ERNIE that reaches SOTA accuracy in content recognition across text, tables, formulas, charts & handwriting, and it supports 109 languages! 🌏 Check it out, play with it, and drop us a ⭐️/❤️ if it helps your work... That means a lot to us🥹 GitHub 👉🏻github.com/PaddlePaddle/Padd… HuggingFace 👉🏻huggingface.co/PaddlePaddle/… AI Studio 👉🏻paddleocr.ai/latest/en/index…

TOP IN TRENDING, TOP IN ACCURACY. Try out the best OCR you can have 👉huggingface.co/spaces/Paddle…

ERNIE 5.0 Preview is making waves in the Text Arena! 💥 Onward and upward. 🚀 #ERNIE #BaiduAI #TextArena

This is a great achievement from the SGLang team! I got immediately asked by a few people what vLLM's plan is. Multimodal generation with omni models is something we have been working on with the community and model vendors. As a sneak peek, Hunyuan-image 3.0 (that beats nano-banana!) already released their official implementation based on vLLM! github.com/Tencent-Hunyuan/H… Personally I’ve always been bullish on omni models - that’s why we’re building a framework that’s diffusion-backend agnostic while fully harnessing vLLM’s efficiency as a core design principle. Stay tuned!

Here is Gemini CLI’s October 27th weekly update for v0.12.0 - 🎉 New Partner Extensions: - 🤗 Hugging Face extension: Access the Hugging Face hub. gemini extensions install github.com/huggingface/hf-mc… - Monday.com extension: Analyze your sprints, update your task boards, etc. gemini extensions install github.com/mondaycom/mcp - Data Commons extension: Query public datasets or ground responses on data from Data Commons. gemini extensions install github.com/gemini-cli-extens… - Model Selection: Choose the Gemini model for your session with /model. (abhipatel12) - Model Routing: Gemini CLI will now intelligently pick the best model for the task. Simple queries will be sent to Flash while complex analytical or creative tasks will still use the power of Pro. This ensures your quota will last for a longer period of time. You can always opt-out of this via /model. (abhipatel12) - Discussion: github.com/google-gemini/gem… - Codebase Investigator Subagent: We now have a new built-in subagent that will explore your workspace and resolve relevant information to improve overall performance. (abhipatel12 and silviojr) - Enable, disable, or limit turns in /settings, plus advanced configs in settings.json (silviojr) - Explore Extension: Users can now open the extensions page in their default browser directly from the CLI using the /extensions explore command. (JayadityaGit) - Configurable Compression: Users can modify the compression threshold in /settings. The default has been made more proactive (scidomino). - API Key Authentication: Users can now securely enter and store their Gemini API key via a new dialog, eliminating the need for environment variables and repeated entry. (galz10) - Sequential Approval: Users can now approve multiple tool calls sequentially during execution. (joshualitt) github.com/google-gemini/gem… 🧵

🚨 Deepseek moment for AI Video. LTX-2 now ranks #3 globally in AI video, built for <$100M while Big Tech burned billions. Efficiency is beating infinite compute.

Introducing ADK Go, a new open-source toolkit designed to empower Go developers to build powerful AI agents. Highlights: 0:00 Introducing the Go ADK 0:53 Build your first Go agent 1:49 Run your agent 2:23 Understanding Tools 3:28 Run your agent with tools 5:00 Multi-agent patterns in Go 7:58 State, Sessions, and Events

🚨Leaderboard Update! A new challenger has stormed the Text Arena, the most competitive field in AI. ERNIE-5.0-Preview-1022 by Baidu @ernieforDevs has reached a score of 1432! The ranking is still preliminary, and we’ll see how it converges. 🔥Huge congrats to the Baidu team on this incredible feat. The competition continues to be fierce!

Join us next Monday, November 10 at 10:00 pm UTC+8 / 9:00 am UTC-5 for the livestream reveal of MeDo @Medo_CodeFree — our AI-powered no-code application builder. Get a first look at how MeDo empowers anyone to create complete, functional apps.

🎉We’re thrilled to share that the ERNIE-5.0-Preview-1022 now ranks #2 (tied) globally on the LMArena Text leaderboard — one of the world’s most recognized benchmarks for large language models driven by real-world use. As part of our commitment, we plan to officially release ERNIE-5.0-Preview-1022 in the near future. This will allow more users and developers to experience and evaluate the model firsthand. Stay tuned — we’ll be sharing it on👉ernie.baidu.com soon!

ERNIE-5.0-Preview-1022, ready for testing in LMArena!