VAGPE MediaBook a call

AI PulseSource

MarkTechPost

45 stories from MarkTechPost that made it into AI Pulse, each rewritten in plain language. From 1 Oct 2026 to 7 Nov 2026, newest first. Looking for something specific? Search every story.

November 2026

  1. In: New AI tools and model updates you can try

    Zyphra ZYFA: FP8 FlashAttention for AMD MI355X, Up to 1.76x Faster

    Zyphra ZYFA adds FP8 FlashAttention for AMD GPUs, up to 1.76× faster inference

    marktechpost.com

  2. In: New AI tools and model updates you can try

    Thinking Machines Releases Inkling 2 Small, a 594B MoE With 17B Active

    Inkling 2 Small, a 594B MoE model with 17B active parameters now available

    marktechpost.com

  3. In: AI tools for video and daily tasks

    Google Releases Gemini Omni 1.2 Flash: 15-Second AI Video at $0.07 per Second

    New Gemini Omni 1.2 Flash makes 15‑sec AI videos for $0.07/second

    marktechpost.com

  4. In: AI tools for video and daily tasks

    Baidu Releases ERNIE 5.5 Flash Base: A 280B MoE Base Model With 14B Active Parameters

    Baidu opens ERNIE 5.5 Flash Base, a 280B MoE model with 14B active params

    marktechpost.com

  5. In: AI tools for video and daily tasks

    Together AI Releases Together Compiler: Agents Make Inference Up to 2.6x Faster

    Together AI’s Compiler speeds inference up to 2.6× for AI agents

    marktechpost.com

  6. In: ChatGPT GIFs and New Vision Model for Creators

    Ricursive AI Releases Ripple-65B Preview: A 65B Dense Open-Weight Model Scoring 80.5 on Terminal-Bench 2.1

    Ricursive Intelligence released Ripple-65B Preview, a 65‑billion‑parameter open‑weight LLM under Apache 2.0 that scores 80.5 on Terminal‑Bench 2.1, offering a high‑capacity model for developers.

    marktechpost.com

  7. In: ChatGPT GIFs and New Vision Model for Creators

    Tencent WeVisOne-VL Unifies Detection, OCR, Segmentation and Matting

    Tencent’s WeVisOne‑VL vision‑language model, available in 2 B and 4 B sizes under Apache 2.0, combines detection, OCR, counting and matting in a single package for image‑heavy apps.

    marktechpost.com

  8. In: New AI tools for everyday use

    webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500

    webAI’s TwIL-8B 8.8B reasoning model now available, free tier for testing

    marktechpost.com

  9. In: New AI tools for quick video & audio work

    Anthropic Releases FlashProve: Cryptographic Proofs That a Reply Came From Committed Weights

    FlashProve offers cryptographic proof that an AI reply came from a specific model

    marktechpost.com

  10. In: New AI tools for quick video & audio work

    NVIDIA Shows 3 Ways to Speed Up MoE Inference on DGX Spark, Lifting Ling-3.0-flash Throughput 42.5%

    NVIDIA shows three tricks to boost MoE inference speed by up to 42% on a single DGX

    marktechpost.com

  11. In: Practical AI tools you can try now

    NVIDIA Releases Auto Ontology: An Open Semantic Layer That Grounds AI Agent SQL

    NVIDIA Auto Ontology adds a plug‑in for natural‑language data queries, no code needed

    marktechpost.com

  12. In: Practical AI tools you can try now

    The Letter Company Releases Letterpress-1: A Coding Agent Built for Growth Engineering

    Letterpress‑1 coding agent offers a drag‑drop UI to build simple growth‑hacks

    marktechpost.com

  13. In: New AI tools for creators and makers

    DeepSeek Open-Sources dsh-superagent, a Router-Worker Coding Agent

    DeepSeek releases dsh‑superagent, a router‑worker AI that writes code snippets on demand

    marktechpost.com

  14. In: New AI tools for creators and makers

    Nace AI Open-Sources Drex Max v1: A 27B Decision Model That Runs Locally on Apple Silicon

    Drex Max v1, a 27B decision model runs locally on Apple Silicon Macs – free download

    marktechpost.com

  15. In: New AI tools for creators and makers

    AI Security Agents Compared: Claude, Codex, Mantis, Cloudflare

    AI Security Agents showdown: Claude, Codex, Mantis, Cloudflare – quick guide to pick one

    marktechpost.com

  16. In: New AI tools for creators and freelancers

    Meet Factory Droid 1.0: A Model-Independent Coding Agent Harness Built on Script Mode

    Factory Droid 1.0 lets you run AI coding agents without writing code, free preview now

    marktechpost.com

  17. In: New AI tools for creators and freelancers

    Vercel Labs Open-Sources agent-desktop for Agent Computer Use

    Vercel’s agent-desktop gives any AI a desktop view, install via Rust CLI today

    marktechpost.com

October 2026

  1. In: New AI tools you can try today

    Meet TokenEconomizer: Microsoft’s Plugin That Cuts Coding Agent Costs by Up to 30.9%

    Try TokenEconomizer: open‑source plug‑in that can cut coding‑agent costs up to 30.9%

    marktechpost.com

  2. In: New AI tools you can try today

    Samaya AI Releases Sterling: A 3B-Active Finance Research Agent Model that Beats Claude Fable 5.1 on FrontierFinance at 11x Lower Cost

    Samaya’s Sterling model delivers finance research answers at 11× lower cost

    marktechpost.com

  3. In: Practical AI tools you can start using

    Baseten Releases joey-1, a 2B Decision Model at $0.01 per 1M Tokens

    Baseten’s Joey‑1 2B decision model runs at $0.01 per 1M tokens – cheap API

    marktechpost.com

  4. In: Free multilingual transcription model released

    vLLM Semantic Router Releases d3-ultra: A 27B Open Multimodal Decision Model

    vLLM Semantic Router unveiled d3‑ultra, a 27 billion‑parameter multimodal decision model that reads text, images and video once and returns probability scores, enabling richer AI reasoning.

    marktechpost.com

  5. In: Google's new Gemini and a free transcriber

    MiniMax made the AI agent skills its own coding team uses every day free and open source

    marktechpost.com

  6. In: Anthropic files to go public

    OpenAI launched GPT-6.1 Terra, a fast model at half the API price of GPT-6.1 Sol

    marktechpost.com

  7. In: Free AI scanners and an AI price war

    Anthropic launched OSS Scanner, a free AI tool that checks key open-source code for security holes

    marktechpost.com

  8. In: OpenAI takes its AI off the live internet

    Nvidia released a small open 4 GB AI model for voice agents and game characters on your own device

    marktechpost.com

  9. In: One prompt line that makes AI more honest

    DeepSeek released a new version of its V4.1-Flash model and opened a search API

    marktechpost.com

  10. In: AI music now comes with a label

    SpaceXAI released Grok Voice 3.0, a voice AI model ranked first on Artificial Analysis

    marktechpost.com

  11. In: Meta's new model and AI tools for work

    12 AI live translation tools compared on speed, languages and price, including Google Live Translate

    marktechpost.com

  12. In: New AI tools for quick content creation

    Try Alibaba's Qwen-Image 2.1 Turbo, a fast 7B image generator you can test free

    marktechpost.com

  13. In: New AI tools for quick content creation

    OpenAI Decisions API enters beta, returns typed answers 10x faster for quick queries

    marktechpost.com

  14. In: New AI tools for quick content creation

    Saluki 27B 2-bit model offers cheap tool-calling power, runs on modest hardware

    marktechpost.com

  15. In: New AI agents and security tools for faster workflows

    Architect's Liquid Inference lets you buy LLM power by the second

    marktechpost.com

  16. In: New AI agents and security tools for faster workflows

    NVIDIA PivotOPD helps AI agents correct mistakes during multi-turn chats

    marktechpost.com

  17. In: Practical AI tools for content creators

    Mistral Large 4, a 1.05T multimodal model now in preview, can power advanced video scripts

    marktechpost.com

  18. In: Practical AI tools for content creators

    Rho-1 by Reka, a 19B model creates short video clips from text, useful for faceless YouTube

    marktechpost.com

  19. In: Free AI tools and new open models

    Meet Together Link: A Free CLI That Runs Open Models Like Kimi K3 and GLM 5.3 Inside Claude Code, Codex, and OpenCode

    marktechpost.com

  20. In: Free AI tools and new open models

    Reflection AI Introduces Beam: A 501B Open-Weight MoE Model With 23B Active Parameters for Coding and Agentic Workloads

    marktechpost.com

  21. In: New AI tools and model updates you can try

    Aleph Alpha Releases Kolibri: A 78.1B Open-Weight English-German MoE Model With Only 3.46B Active Parameters

    Kolibri offers a 78B multilingual model you can access via free API

    marktechpost.com

  22. In: New AI tools and model updates you can try

    Google Research Moves Federated Learning Into TEEs: Gboard Now Trains With Externally Verifiable Differential Privacy

    Google’s Gboard now trains with verifiable differential privacy on‑device

    marktechpost.com

  23. In: New AI tools for everyday creators

    IBM Bob self-hosted lets you generate code without sending data online

    marktechpost.com

  24. In: New AI tools for everyday creators

    Inside NVIDIA's IsaacTeleop: From Hand and Controller Tracking to Robot Actions with the Graph-Based…

    marktechpost.com

  25. In: New AI tools for quick content creation

    AWS Strands Decider 2B picks options in ~115 ms, useful for quick content decisions

    marktechpost.com

  26. In: New AI tools and model updates you can try

    Perplexity's Photon engine speeds search results to 65 ms latency

    marktechpost.com

  27. In: New AI tools and model updates you can try

    OpenAI GPT-6.1 Sol offers coding help at one-fifth Astra's price

    marktechpost.com

  28. In: New AI tools and model updates you can try

    Google Gemini 4 Argon now supports 1 million output tokens for longer tasks

    marktechpost.com

Get AI Pulse

Free, two ways. Pick one, or both.

Or join our free Friday webinar