Results for “context engineering”

…... Codeium (Windsurf) ▷ #announcements (2 messages): AI Engineering Summit Tickets, Windsurf Wave 3 Features, Model Context Protocol, Customizable App Icons, Turbo Mode Enhancements Win Free Tickets to the AI Engineering Summit!: We're…

…with extensive long-context pretraining (~630B tokens for 32k context, ~209B for 128k), FP8 training, and a large MoE expert count (~37B). Benchmarks show competitive performance with notable improvements in SWE-Bench and other reasoning…

…reactions to a very successful GPT 5.5 model launch. But in context they mark a very noteworthy reaction that you, dear reader, should probably be alerted to if you aren’t already taking this extremely seriously. The proximal trigger for…

…of Large Language Models: Typically, training LLMs with long context sizes is computationally expensive, requiring extensive training hours and GPU resources. Existing long-context extension methods usually need additional tra... ▷ #general…

…model with 40B parameters, over 20T tokens, with MLA for long context efficiency. Hugging Face MCP Course: @reach_vb announced Hugging Face has dropped the MCP Course, covering everything you need to know about Model Context Protocol and…

…multimodal dense model under Apache 2.0, with 262K native context extendable to 1M via YaRN, while also highlighting the already-released Qwen3.8-2.4T-A95B max-tier model (announcement, perf thread). The 27B model is notable because it is…

…& Updates Mistral AI released a new 7B 0.2 base model with 32k context window, announced at a hackathon (13k views) Quantized 4-bit Mistral 7B models released, enabling 2x faster inference with 70% less VRAM using QLoRA finetuning (14k…

…(Apache 2.0), and support for 119 languages with a 32K–128K context. Initial results show Qwen3 models are competitive for their size class, with the 253B-A22B model approaching DeepSeek R1’s GPQA score @ArtificialAnlys. They also…

…the new gpt2-chatbot on lmsys, and extending Llama-3-8B to 1m context, but otherwise no clear top story emerges. You can check out the WebSim/WorldSim podcast as Nous Research gets ready to relaunch it after briefly taking it down due to…

…the founders' partnership, calling their story remarkable. RAG, Context Engineering, and Data Quality: The term Context Rot was highlighted as an excellent and useful term by @jxmnop. DeepLearningAI provided a technical breakdown of how…

…tools. Codeium (Windsurf) ▷ #announcements (2 messages): AI Engineering Summit Tickets, Windsurf Wave 3 Features, Model Context Protocol, Customizable App Icons, Turbo Mode Win Tickets to AI Engineering Summit!: We’re giving away 3 tickets …

…capturing equations across LaTeX, MathML, , inline, and image contexts—addressing coverage gaps in typical parsers. NVIDIA reports marked gains on math and code tasks after adding it. Details from @KarimiRabeeh and @ctnzr; commentary by…