Salesforce Koa Is a CRM Reasoning Model Built on NVIDIA Nemotron
Salesforce Koa is a CRM reasoning model built on NVIDIA Nemotron 3 Super. It is in a select pilot, with US availability expected in…
Author: Abdessalam Alaoui
I'm Abdessalam Alaoui, founder of Musthave.AI. I write practical guides to AI models, tools, generative media, and ways real people make money with this stuff. No press-release fluff. 20+ years shipping digital products; you'll also find my work around HowTik, Yumasty, and IncomeHelpful.
Salesforce Launches AIforce for Claude, Slack and Governed CRM Actions
Salesforce AIforce brings CRM data, workflows and governed actions into Claude, Slack and other AI interfaces. The Claude integration is in beta.
TypeSafe Launches Jev, an AI Model for Typed Decisions Instead of Text
TypeSafe's Jev model returns typed decisions instead of generated text, now with native Vercel AI Gateway support. See API paths, pricing, speed and limits.
Qwen Publishes a New Studio Overview for Its Multimodal AI Assistant
Qwen Studio brings reasoning, multimodal input, generation, code and Artifacts to web, mobile and desktop. Here is what the overview confirms.
InclusionAI Releases LLaDA-UI, a 16.7B Diffusion Model for GUI Agents
LLaDA-UI is a 16.7B diffusion vision-language model that turns native-resolution interface images into structured GUI actions.
Ling-3.0-flash-VL Launches as Its Benchmark Score Changes From 42 to 25
Ling-3.0-flash-VL has 124B total parameters, 5.5B active parameters and two benchmark scores tied to different index versions.
Tencent Open-Sources AuK for Speech Generation, Editing and Separation
Tencent AuK is a 1.5B open speech model for generation, editing, enhancement and separation. AuK-Flash uses four inference steps.
OpenBMB Releases MiniCPM5-2B for Local AI With 131K Context
MiniCPM5-2B is a dense local AI model with 131K context, GGUF and MLX builds, open data and a sampling warning developers should know.
Z.ai Launches GLM-5.3-Flash With 18B Active Parameters and 1M Context
GLM-5.3-Flash combines 320B total parameters, 18B active parameters, native multimodality and 1M context. Here is what that means for deployment.
Visual Studio 2026 BYOK Adds OpenAI, Anthropic, Ollama and Foundry to Agent Mode
Visual Studio 2026 BYOK now connects OpenAI, Anthropic, Ollama and Foundry to a new Copilot SDK agent, with breaking migration and governance changes.
Google Agent Search Lets Developers Disable Paid AI Add-ons Per Query
Google Agent Search now exposes per-query switches for semantic search, personalization and generative answers. Applicable add-ons otherwise default on.
IBM Separates Shallow Quantum Circuits and Restricted LLMs on Theoretical Tasks
IBM researchers proved theoretical separations between shallow quantum circuits and restricted LLMs. Here is what the result does and does not show.