Ready-To-Use Watch Template

Open-Source AI Model Releases & Benchmarks

Stay ahead of rapid open-weight AI developments. Get a focused weekly summary of newly released model weights, MMLU/Arena benchmarks, and Hugging Face leaderboards.

App Store (iOS) 📱 Google Play (Android)
⏱️ Checks Weekly on Monday ✓ Hugging Face & ArXiv 🤖 Gemini AI Filtered

What you receive in the app:

THE SIGNAL (Gemini AI Takeaway)
  • DeepSeek-V3 & R1: Published open weights matching commercial reasoning benchmarks on AIME and MATH.
  • Llama 3.3 70B: Community evaluations confirm 405B level quality at high token/sec inference throughput.
  • Quantization: GGUF and EXL2 formats uploaded for local consumer hardware deployment via Ollama and LM Studio.
Direct Citations: Hugging Face, ArXiv AI, LMSYS Chatbot Arena.