All tags

Topic: "context-windows"

    Reasoning Price War 2: Mistral Magistral + o3's 80% price cut + o3-pro
    not much happened today
    Google's Agent2Agent Protocol (A2A)
    DeepCoder: A Fully Open-Source 14B Coder at O3-mini Level
    OpenAI adopts MCP
    not much happened today
    Cohere's Command A claims #3 open model spot (after DeepSeek and Gemma)
    not much happened today
    DeepSeek's Open Source Stack
    not much happened today
    GPT 4.5 — Chonky Orion ships!
    lots of small launches
    not much happened today
    Gemini 2.0 Flash GA, with new Flash Lite, 2.0 Pro, and Flash Thinking
    Bespoke-Stratos + Sky-T1: The Vicuna+Alpaca moment for reasoning
    not much happened today
    Perplexity starts Shopping for you
    Did Nvidia's Nemotron 70B train on test?
    Not much technical happened today
    not much happened today
    Llama 3.2: On-device 1B/3B, and Multimodal 11B/90B (with AI2 Molmo kicker)
    Replit Agent - How did everybody beat Devin to market?
    Ideogram 2 + Berkeley Function Calling Leaderboard V2
    not much happened today
    GPT4o August + 100% Structured Outputs for All (GPT4o August edition)
    AlphaProof + AlphaGeometry2 reach 1 point short of IMO Gold
    Mistral Large 2 + RIP Mistral 7B, 8x7B, 8x22B
    Llama 3.1 Leaks: big bumps to 8B, minor bumps to 70b, and SOTA OSS 405b model
    DataComp-LM: the best open-data 7B model/benchmark/dataset
    Mini, Nemo, Turbo, Lite - Smol models go brrr (GPT4o-mini version)
    Mini, Nemo, Turbo, Lite - Smol models go brrr (GPT4o version)
    Qdrant's BM42: "Please don't trust us"
    GraphRAG: The Marriage of Knowledge Graphs and RAG
    Mozilla's AI Second Act
    Nemotron-4-340B: NVIDIA's new large open models, built on syndata, great for syndata
    Cursor reaches >1000 tok/s finetuning Llama3-70b for fast file editing
    Not much happened today
    OpenAI's Instruction Hierarchy for the LLM OS
    Llama-3-70b is GPT-4-level Open Model
    Cohere Command R+, Anthropic Claude Tool Use, OpenAI Finetuning
    Evals-based AI Engineering
    Jamba: Mixture of Architectures dethrones Mixtral
    AI gets Memory
    GPT4Turbo A/B Test: gpt-4-0125-preview
    Is Google's Gemini... legit?